Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis

Modern Supramolecular Chemistry Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Related Titles Sc...

Author: Francois Diederich | Peter Stang | Rik R. Tykwinski

286 downloads 2190 Views 10MB Size Report

This content was uploaded by our users and we assume good faith they have the permission to share this book. If you own the copyright to this book and it is wrongfully on our website, we offer a simple DMCA procedure to remove your content from our site. Start by pressing the button below!

Report copyright / DMCA form

DOWNLOAD PDF

Modern Supramolecular Chemistry

Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski

Related Titles Schalley, C. A. (ed.)

Analytical Methods in Supramolecular Chemistry 2007 ISBN: 978-3-527-31505-5

Dodziuk, H. (ed.)

Cyclodextrins and Their Complexes Chemistry, Analytical Methods, Applications 2006 ISBN: 978-3-527-31280-1

Cragg, P.

A Practical Guide to Supramolecular Chemistry 2005 ISBN: 978-0-470-86654-2

Gleiter, R., Hopf, H. (eds.)

Modern Cyclophane Chemistry 2004 ISBN: 978-3-527-30713-5

Modern Supramolecular Chemistry Strategies for Macrocycle Synthesis

Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski

The Editors Prof. Dr. François Diederich ETH Hönggerberg, HCI Laboratory for Organic Chemistry Wolfgang-Pauli-Str. 10 8093 Zürich Switzerland Prof. Dr. Peter J. Stang University of Utah Department of Chemistry Henry Eyring Building Salt Lake City, UT 84112 USA Prof. Dr. Rik R. Tykwinski University of Alberta Department of Chemistry Edmonton, Alberta, T6G 2G2 Canada

All books published by Wiley-VCH are carefully produced. Nevertheless, authors, editors, and publisher do not warrant the information contained in these books, including this book, to be free of errors. Readers are advised to keep in mind that statements, data, illustrations, procedural details or other items may inadvertently be inaccurate. Library of Congress Card No.: applied for British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library. Bibliographic information published by the Deutsche Nationalbibliothek Die Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliograﬁe; detailed bibliographic data are available in the Internet at http://dnb.d-nb.de. # 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim All rights reserved (including those of translation into other languages). No part of this book may be reproduced in any form – by photoprinting, microﬁlm, or any other means – nor transmitted or translated into a machine language without written permission from the publishers. Registered names, trademarks, etc. used in this book, even when not speciﬁcally marked as such, are not to be considered unprotected by law. Typesetting Thomson Digital, India Printing Strauss GmbH, Mörlenbach Binding Litges & Dopf GmbH, Heppenheim Printed in the Federal Republic of Germany Printed on acid-free paper ISBN: 978-3-527-31826-1

V

Contents Preface XIII List of Contributors

1 1.1 1.2 1.2.1 1.2.2 1.2.3 1.2.4 1.2.5 1.2.6 1.3 1.4 1.4.1 1.4.2

2 2.1 2.2 2.2.1 2.2.2 2.2.3 2.2.3.1 2.2.3.2

XV

Bioactive Macrocyclic Peptides and Peptide Mimics 1 Rob M.J. Liskamp, Dirk T.S. Rijkers, and Saskia E. Bakker Introduction 1 Selected Cyclic Peptides 4 Vancomycin 4 Lantibiotic: Nisin 6 Cyclosporin A 10 Cyclotheonamide A and B 13 cyclo RGD Peptides as aVb3 Antagonists 16 SH2 Domain-Binding Peptides 19 Conclusions 22 Experimental: Selected Procedures 22 Synthesis of Bicyclic Peptide 9: an Alkene-bridged Mimic of the Vancomycin C-D-E Cavity 22 Synthesis of Cyclic Peptide 14: an Alkyne-bridged Mimic of the Nisin A-Ring Fragment 22 References 25 Macrocycles by Ring-closure Metathesis 29 Joëlle Prunet, Anderson Rouge dos Santos, and Jean-Pierre Férézou Introduction 29 How to Cyclize? 32 A Thermodynamic versus Kinetic Issue 32 General Experimental Conditions 34 Inﬂuence of Polar Complexing Groups 35 A Decisive Factor for Success 35 The Titanium Trick 36

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

VI

Contents

2.2.3.3 2.2.4 2.2.5 2.2.5.1 2.2.5.2 2.3 2.3.1 2.3.2 2.3.3 2.3.4 2.4 2.5 2.5.1 2.5.2 2.5.3 2.5.4 2.5.5 2.5.6 2.6 2.6.1 2.6.1.1 2.6.1.2 2.6.1.3 2.6.1.4 2.6.1.5 2.6.1.6 2.6.2 2.6.3 2.6.4 2.7 2.8 2.8.1 2.8.2 2.8.3 2.8.4 2.8.5 2.8.6

A Particularly Favorable Case: The Template-Directed RCM 37 Chemoselectivity 37 Substrate ‘‘Tuning’’ 38 Conﬁgurational/Conformational Aspect 38 Inﬂuence of Functional Group Protection 39 Factors Inﬂuencing the Double-Bond Conﬁguration 39 A Thermodynamic versus Kinetic Issue 40 General Experimental Conditions 42 Substrate ‘‘Tuning’’ 42 Solutions 44 Ene-yne M-RCM 45 Tandem Processes Involving M-RCM 46 Tandem CM/RCM 47 Tandem ROM/RCM 47 Tandem RCM/ROM/RCM 48 Ring-Expansion Metathesis 48 Other One-Pot Multistep Processes 49 M-RCM as Part of MCR Strategies 50 Representative Synthetic Applications 50 Salicylihalamides/Oximidines: Potent Antitumor Agents with Selective anti-V-ATPase Activity 51 Salicylihalamides: Inﬂuence of the Catalyst 52 Salicylihalamides: Inﬂuence of the Phenol Protecting Group 53 Salicylihalamides: Inﬂuence of the Alcohol-Protecting Group at C13 54 Salicylihalamides: Inﬂuence of the Nature of the C15 Side Chain 54 Salicylihalamides: Miscellaneous 55 Oximidines: Use of Relay Ring-closing Metathesis 55 Radicicol-Type Macrolides: a Promising Family of Anticancer Resorcylides 56 Coleophomones: a Versatile Access to this Class of Complex Polycyclic Diterpenes 58 RCM in Supramolecular Chemistry 59 Conclusion and Perspectives 62 Experimental: Selected Procedures 62 Synthesis of Compound 3 with Catalyst S1 62 Synthesis of Compound 6 with Catalyst G1 62 Synthesis of Compound 8 with Catalyst G2 63 Synthesis of Compound 16 (R¼SiMe2tBu) with Catalyst G1/Ti(OiPr)4 63 Synthesis of Compound 49 by RCAM 63 Synthesis of Compound 53 with G1 by ene-yne RCM 63 References 64

Contents

3 3.1 3.2 3.3 3.4 3.5 3.5.1 3.5.2

4 4.1 4.1.1 4.1.2 4.2 4.2.1 4.2.2 4.2.2.1 4.2.2.2 4.3 4.3.1 4.3.2 4.3.3 4.3.4 4.3.5 4.3.6 4.4 4.4.1 4.4.1.1 4.4.1.2 4.4.1.3 4.4.1.4 4.4.2

Supramolecular Macrocycle Synthesis by H-bonding Assembly Pablo Ballester and Javier de Mendoza Introduction 69 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds 74 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds 90 Summary and Outlook 105 Experimental: Selected Procedures 106 Solid State Formation of the Hexameric Capsule Derived from Pyrogallol[4]arene (50c) 106 Crystals of the Host-Guest Arrangement of 52@(50b)6 106 References 108 Cucurbit[n]urils 113 Wei-Hao Huang, Simin Liu, and Lyle Isaacs Introduction 113 Synthesis and Structure of Cucurbit[6]uril and Decamethylcucurbit[5]uril 113 Molecular Recognition Properties of Cucurbit[6]uril 114 New Members of the Cucurbit[n]uril Family 115 Proposed Mechanism of Cucurbit[n]uril Formation 115 Synthesis and Structure of Cucurbit[n]uril Homologs (n ¼ 5, 7, 8, 10) 116 Reaction Conducted Under Milder Conditions 116 CB[5] Can be Released from CB[10].CB[5] to Yield Free Cucurbit[10]uril 117 Applications of Members of the Cucurbit[n]uril Family 118 Preparation of Molecular Switches 118 Self-Assembled Dendrimers 119 Preparation of Molecular Machines 119 Preparation of Complex Self-Sorting Systems 121 Allosteric Control of the Conformation of a Calix[4]arene Inside CB[10] 122 As a Carrier of Anti-Cancer Agents 123 Experimental Support for the Proposed Mechanism of CB[n] Formation 124 S-shaped and C-shaped Methylene-bridged Glycoluril Dimers Synthesis of Methylene-bridged Glycoluril Dimers 124 S- to C-shaped Isomerization of Methylene-bridged Glycoluril Dimers 126 Mechanism of S- to C-shaped Isomerization 126 Implications for the Synthesis of Cucurbit[n]uril Analogs and Derivatives 128 Building-Block Approach to Cucurbit[n]uril Analogs 128

69

124

VII

VIII

Contents

4.4.3 4.4.4 4.5 4.5.1 4.5.2 4.5.3 4.6 4.6.1 4.6.2 4.7 4.8 4.8.1 4.8.2 4.8.3

Building-Block Approach to Cucurbit[n]uril Derivatives 129 Identiﬁcation and Isolation of Inverted Cucurbit[n]urils (n ¼ 6, 7) 130 Direct Functionalization of Cucurbit[n]urils 131 Perhydroxylation and Further Derivatization of CB[5]–CB[8] 131 Multivalent Binding of Sugar-Decorated Vesicles to Lectins 132 Cucurbit[n]uril-based Artiﬁcial Ion Channels 132 Nor-Seco-Cucurbit[10]uril 133 Detection and Isolation of Nor-Seco-Cucurbit[10]uril 134 Molecular Recognition Properties of Nor-Seco-Cucurbit[10]uril 134 Summary and Conclusions 135 Experimental: Selected Procedures 137 Synthesis of Glycolurils 137 Synthesis and Separation of Cucurbit[n]urils 138 Synthesis of Nor-Seco-Cucurbit[10]uril 140 References 141

5

Tetra-urea Calix[4]arenes – From Dimeric Capsules to Novel Catenanes and Rotaxanes 143 Ganna Podoprygorina and Volker Böhmer Introduction 143 Basics of Tetra-urea Calix[4]arenes 148 Synthesis 148 Dimeric Capsules 149 Heterodimers 151 Symmetry Properties 152 Preorganization in Dimers of Tetra-urea Calix[4]arenes 153 General Considerations 153 First Attempts 154 Multimacrocycles 155 Template Synthesis of Multimacrocyclic Calix[4]arenes 155 Double Template Synthesis of Giant Macrocycles 160 Multiple Catenanes Based on Calix[4]arenes 162 General Considerations 162 Bis[2]catenanes 163 Towards Novel Topologies 166 Bis[3]catenanes and Cyclic [8]Catenanes 168 Multiple Rotaxanes 170 Self-sorting and Formation of Larger Assemblies 172 Conclusions and Outlook 176 Experimental: Selected Procedures 177 Synthesis of Tetra-urea 5 (Y ¼ C5H11; m ¼ 9) 177 Synthesis of Bisloop Tetra-urea 8 (Y ¼ C5H11; n ¼ 20) 177 Synthesis of Bis[2]catenane 12 (Y ¼ C5H11; n ¼ 20) 177 Synthesis of Tetra-urea 6a (Y ¼ C5H11; m ¼ 6) 178 Synthesis of Tetraloop Tetra-urea 9 (Y ¼ C5H11; n ¼ 14) 178

5.1 5.2 5.2.1 5.2.2 5.2.3 5.2.4 5.3 5.3.1 5.3.2 5.4 5.4.1 5.4.2 5.5 5.5.1 5.5.2 5.5.3 5.5.4 5.6 5.7 5.8 5.9 5.9.1 5.9.2 5.9.3 5.9.4 5.9.5

Contents

5.9.6 6 6.1 6.1.1 6.1.2 6.2 6.2.1 6.2.2 6.2.3 6.2.3.1 6.2.3.2 6.2.3.3 6.3 6.4 6.4.1 6.4.2 6.4.3 6.4.4 6.4.5 6.4.6 6.4.7

7

7.1 7.2 7.2.1 7.2.2 7.2.2.1 7.2.2.2 7.2.3 7.2.4 7.2.5 7.3 7.4 7.4.1 7.4.2 7.4.3 7.4.4 7.4.5

Synthesis of [8]Catenane 14 (Y ¼ C5H11; n ¼ 14) 179 References 180 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding 185 Amber L. Sadowy and Rik R. Tykwinski Introduction 185 SPM Synthesis through Intermolecular Reactions 186 SPM Synthesis through Intramolecular Reactions 190 Supramolecular SPMs 194 SPMs as Components in Supramolecular Assemblies 195 SPMs in Host–Guest Systems 203 Aggregation and Surface Chemistry of SPMs 208 Aggregation of SPMs 209 Liquid-Crystalline SPMs 215 Adsorption of SPMs on Surfaces 218 Conclusions 224 Experimental: Selected Procedures 224 SPM 13: Pd-Catalyzed Cadiot–Chodkiewicz Conditions 224 SPM 19: Use of Aryltriazene as a Masking Group for Aryl Iodides 224 SPM 20: Eglinton Conditions 225 SPM 33: Hay Conditions 225 Pre-Catenane 56: Breslow Conditions 225 SPM 91: Schiff-base Condensation Conditions 226 Large-scale Synthesis of SPM 124 via an Alkyne Metathesis 226 References 228 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles 233 Laura Pirondini and Enrico Dalcanale Introduction 233 Coordination Cages 234 Dimeric Calixarene-based Coordination Cages 235 Cavitand-based Dimeric Coordination Cages 236 The Apical Functionalization Approach 236 Introduction of the Ligands as Bridging Units 248 Trimeric, Tetrameric, and Hexameric Coordination Cages 252 Self-assembly of Coordination Cages on Surfaces 255 Self-assembly of Multitopic Macrocyclic Complexes 263 Conclusion 271 Experimental: Selected Procedures 272 Tetrapicolyl-bridged Cavitand 31a 272 Tetracyano Cavitand 40 272 AC-dibridged Tolylpyridyl Cavitand 35 272 fac-Br(CO)3Re AC Ditopic Cavitand Complex 36 273 Tetratopic Cavitand Complex 48 273 References 274

IX

X

Contents

8

8.1 8.2 8.2.1 8.2.2 8.3 8.3.1 8.3.2 8.3.3 8.3.4 8.3.5 8.3.6 8.3.7 8.4 8.4.1 8.4.2 8.5 8.5.1 8.5.2 8.5.3 8.5.4 8.5.5 8.5.6 8.5.7 8.6 8.7

9 9.1 9.2 9.2.1 9.2.2 9.2.3 9.2.4 9.2.5 9.3 9.4 9.5 9.5.1 9.5.2

New Properties and Reactions in Self-assembled M6L4 Coordination Cages 277 Makoto Fujita and Michito Yoshizawa Introduction 277 Self-assembly of Hollow Complexes 278 M6L4 Octahedral Complex 280 Large-scale Production of the M6L4 Complex 280 Inclusion Properties 288 Inclusion of Adamantane and Carborane 288 Inclusion Geometry 289 Bimolecular Recognition 291 Recognition of Bulky Guests 293 The Recognition of Azobenzene and Stilbene 295 Molecular Ice 296 Peptide Recognition 297 New Physical Properties 299 Redox Control of Ferrocene 299 Induction of Intermolecular Spin–spin Interaction 299 New Reactions 301 [2þ2] Oleﬁn Photodimerization 302 Pairwise-selective Oleﬁn Photodimerization 303 Unusual [2þ2] Photoaddition 303 Diels-Alder Reaction 303 Alkane Oxidation 306 Wacker Oxidation 306 Discrete Siloxane Synthesis 308 Conclusion 308 Experimental: Synthesis of M6L4 Cage 2 309 References 309 Anion-binding Macrocycles 315 Evgeny A. Katayev, Patricia J. Melfi, and Jonathan L. Sessler Introduction 315 Ditopic Receptors and Receptors for Ion Pairs 317 Crown Complexes 318 Calixarenes 321 Cholapods 325 Pyrroles 326 Miscellaneous 330 Receptors with Different Binding Sites 332 Conclusions 341 Experimental: Selected Procedures 342 Macrocycle H2SO4.53 342 Macrocycle 55 342 References 343

Contents

10 10.1 10.1.1 10.2 10.2.1 10.2.2 10.2.3 10.2.4 10.2.5 10.2.6 10.2.7 10.3 10.4 10.4.1 10.4.2 10.4.3 10.4.4

Rotaxane and Catenane Synthesis 349 James A. Wisner and Barry A. Blight Introduction 349 General Comments 349 Macrocyclization Reactions Resulting in Interlocked Products 351 Williamson Ether Synthesis 351 Quaternization of Aromatic Amines (Menschutkin Reaction) 351 Condensation of Amines with Acid Chlorides 354 Oxidative Acetylide Coupling 361 Alkene Metathesis 366 Imine Formation/Reductive Amination 374 Metal-Ligand Coordination 379 Conclusions 384 Experimental: Selected Procedures 384 [2]Catenane 14 384 [2]Catenane 25 386 [2]Rotaxane 81 386 [2]Catenane 118 386 References 387 Index

393

XI

XIII

Preface The 1987 Nobel Prize in Chemistry was awarded to Donald J. Cram, Jean-Marie Lehn, and Charles J. Pedersen for ‘‘their development and use of molecules with structure-speciﬁc interactions of high selectivity’’. At this time, the award was bestowed for the synthesis of molecules that mimicked important biological processes – what Lehn deemed supramolecular chemistry. In the two decades that have followed, this ﬁeld has expanded greatly, and supramolecular breakthroughs in organic synthesis, molecular electronics, and materials science have now been realized. It is interesting to note, however, that in many respects supramolecular chemistry has remained close to its roots. The pioneering efforts that merited the Nobel Prize were based primarily on the synthesis of macrocycles. These were the crown ethers, the cryptands, the cavitands, and other host molecules that ultimately provide a welcoming and selective environment for a particular guest species, whether it be a neutral molecule, a cation, or an anion. Both historically and in the present day, supramolecular chemistry beautifully marries two scientiﬁc disciplines: organic synthesis and physical organic chemistry. It is here that the most modern aspects of this ﬁeld of chemistry share the spotlight. No longer is the objective simply to mimic biological systems such as enzymes, but rather, today’s supramolecular chemist is limited only by his or her imagination as to the role that a macrocycle might play in some well-orchestrated chemical scheme. The goal of the present monograph is to tie together these seemingly diverse achievements under a common heading: the synthesis of macrocycles for use in supramolecular chemistry. To this end, the biological relevance of macrocycles continues to play a pivotal role, as illustrated by a chapter on macrocyclic peptides by Liskamp and co-workers. The emergence of ring-closing metathesis as a preeminent synthetic strategy for constructing naturally occurring macrocycles is described next by Prunet et al. Synthetic efforts toward macrocycles continue to be inspired by Nature, and these stories are recounted by de Mendoza and Ballester (hydrogen-bonding assembly), Isaacs and coworkers (curcurbiturils), Böhmer and Podoprygorina (tetra-urea calixarenes), and Sessler and coworkers (anion-binding macrocycles). From here, supramolecular chemistry melds into materials chemistry, where shape-persistent macrocycles (Tykwinski and Sadowy), three-dimensional

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

XIV

Preface

architectures (Dalcanale and Pirondini), molecular containers (Fujita and Yoshizawa), and rotaxanes (Wisner and Blight) share the stage. While it is not possible to review comprehensively each of the above topics, we believe this monograph provides expert insight and advice, covering both synthetic endeavors and applications of the resulting products. Most of all, however, we hope that these 10 chapters will instill the inspiration to further expand the boundaries of this captivating ﬁeld of research. This book results from the substantial efforts of a number of people. Most importantly, we appreciate the contributions of the authors. In this era of dwindling time and funding, their efforts and expertise have provided a monograph that is both interesting to read and a scientiﬁc resource. We express our gratitude to Ms. Annie Tykwinski for the original illustration that became the cover of this book, and we would like to thank Drs. Manfred Köhl and Andreas Sendtko at Wiley-VCH, as well as the production staff, for their aid in preparing this monograph. November 2007

François Diederich Peter J. Stang Rik R. Tykwinski Zurich, Salt Lake City, Edmonton

XV

List of Contributors Saskia E. Bakker Utrecht University Faculty of Science Department of Pharmaceutical Sciences Medicinal Chemistry and Chemical Biology P.O. Box 80082 3508 TB Utrecht The Netherlands Pablo Ballester Institute of Chemical Research of Catalonia (ICIQ) and Institució Catalana de Recerca i Estudis Avançats (ICREA) Avda. Països Catalans, 16 43007 Tarragona Spain Barry A. Blight University of Western Ontario Faculty of Science Department of Chemistry 1151 Richmond St., Chemistry Building London, Ontario, N6A 5B7 Canada

Volker Böhmer Johannes Gutenberg-Universität Mainz Fachbereich Chemie, Pharmazie und Geowissenschaften Duesbergweg 10–14 55099 Mainz Germany Enrico Dalcanale Università di Parma Dipartimento di Chimica Organica e Industriale Viale G. Usberti 17/A 43100 Parma Italy Jean-Pierre Férézou École Polytechnique Laboratoire de Synthèse Organique, DCSO CNRS UMR 7652 91128 Palaiseau France

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

XVI

List of Contributors

Makoto Fujita University of Tokyo Department of Applied Chemistry School of Engineering and CREST (Core Research for Evolution Science and Technology) Japan Science and Technology Agency (JST) 7-3-1 Hongo, Bunkyo-ku Tokyo 113-8656 Japan Wei-Hao Huang University of Maryland Department of Chemistry and Biochemistry College Park, MD 20742 USA Lyle Isaacs University of Maryland Department of Chemistry and Biochemistry College Park, MD 20742 USA Evgeny A. Katayev Russian Academy of Sciences A.N. Nesmeyanov Institute of Organoelement Compounds Vavilov St., 28 119899 Moscow Russia

Rob M. J. Liskamp Utrecht University Faculty of Science Department of Pharmaceutical Sciences Medicinal Chemistry and Chemical Biology P.O. Box 80082 3508 TB Utrecht The Netherlands Simin Liu University of Maryland Department of Chemistry and Biochemistry College Park, MD 20742 USA Patricia J. Melfi University of Texas at Austin Department of Chemistry and Biochemistry 1 University Station - A5300 Austin, Texas 78712 USA Javier de Mendoza Institute of Chemical Research of Catalonia (ICIQ) Avda. Països Catalans, 16 43007 Tarragona Spain Laura Pirondini Università di Parma Dipartimento di Chimica Organica e Industriale Viale G. P. Usberti 17/A 43100 Parma Italy

List of Contributors

Ganna Podoprygorina Johannes Gutenberg-Universität Mainz Fachbereich Chemie, Pharmazie und Geowissenschaften Duesbergweg 10–14 55099 Mainz Germany

Jonathan L. Sessler University of Texas at Austin Department of Chemistry and Biochemistry 1 University Station - A5300 Austin, Texas 78712 USA

Joëlle Prunet École Polytechnique Laboratoire de Synthèse Organique, DCSO CNRS UMR 7652 91128 Palaiseau France

Rik R. Tykwinski University of Alberta Department of Chemistry Edmonton, Alberta, T6G 2G2 Canada

Dirk T. S. Rijkers Utrecht University Faculty of Science Department of Pharmaceutical Sciences Medicinal Chemistry and Chemical Biology P.O. Box 80082 3508 TB Utrecht The Netherlands Amber L. Sadowy University of Alberta Department of Chemistry Edmonton, Alberta, T6G 2G2 Canada Anderson Rouge dos Santos Universidade Federal do Rio de Janeiro Instituto de Química Caixa Postal 068534 Ilha do Fundão, CT, Bloco A CEP 21941-972 Rio de Janeiro RJ Brazil

James A. Wisner University of Western Ontario Faculty of Science Department of Chemistry 1151 Richmond St., Chemistry Building London, Ontario, N6A 5B7 Canada Michito Yoshizawa Department of Applied Chemistry School of Engineering The University of Tokyo and CREST (Core Research for Evolution Science and Technology) Japan Science and Technology Agency (JST) 7-3-1 Hongo, Bunkyo-ku, Tokyo 113-8656 Japan

XVII

j1

1 Bioactive Macrocyclic Peptides and Peptide Mimics Rob M.J. Liskamp, Dirk T.S. Rijkers, and Saskia E. Bakker

1.1 Introduction

The number of both naturally occurring and synthesized biologically active cyclic peptides, modiﬁed peptides, and peptide mimics is rapidly increasing. So far, many of these sometimes biologically very potent peptides have unfortunately unknown molecular targets or mechanisms of action. This certainly opens up very interesting and challenging research areas with respect to uncovering these targets and/or molecular mechanisms of biological activity, which – depending on the biological action – can be very promising, for example, for the development of new drugs. However, merely a review of most biologically active cyclic peptides – if this were possible at all – is not the aim of this chapter. Instead, we wish to focus on selected bioactive macrocyclic peptide systems with known molecular peptide or protein targets as well as details about their molecular interaction mechanism. Where possible, we would like to discuss the contribution of the cyclic nature of the peptides to the molecular mechanism of interaction and the ensuing biological activity. We will therefore not include cyclic peptides and mimics merely interacting with membrane lipids or cyclic peptides interacting with DNA or RNA. Each of these topics deserves a review on its own, especially in light of the increasing interest in membrane proteins and transcription activators/regulators. The selection of cyclic peptides interacting with known molecular targets in this chapter is largely determined by their relevance in relation to possible treatments of diseases. In this respect, probably vancomycin and cyclosporin are the most wellknown cyclic peptides containing modiﬁed amino acids, which have had a profound inﬂuence on the treatment of life-threatening diseases. The vancomycin-related antibiotics [1] are outstanding examples of cyclic peptide systems containing multiple knotted side chains by which almost absolute control over the shape of the molecule is achieved, leading to efﬁcient binding of crucial fragments of the cell-wall precursor of disease-causing bacteria.

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

2

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics In the larger peptide antibiotic compounds comprising the class of lantibiotics, the shape of the molecule is determined by several cyclic peptides, including two annulated peptide rings, present within one molecule, giving the lantibiotic a unique way to interact with the target molecule lipid II and subsequent pore-forming capabilities in phospholipid membranes [2]. In an increasing number of examples, the diversity of functions and activities of proteins can be reﬂected in their smaller peptide counterparts, and this is especially prominent in RGD-containing cyclic peptides and peptide mimics derived from the corresponding RGD sequence in proteins. These are capable of interacting with a variety of integrin receptors (aVb3 and aIIbb3) involved in cellular adhesion and migration [3]. As a result, the RGD sequence is, for example, crucial in the construction of (cyclic) peptide compounds used for molecular imaging and treatment of diseases such as cancer and infections. However, it is good to realize that many biologically relevant cyclic peptides are not derived from or do not correspond to particular peptide sequence(s) in a larger protein molecule. These cyclic peptides are produced by micro-organisms, and although they are macrocyclic structures, they are of course much smaller than proteins. In addition to vancomycin and the lantibiotics mentioned above, other important examples are cyclosporin A [4] and cyclotheonamide [5]. The former cyclic peptide is the widely used immunosuppressive drug while the latter is capable of speciﬁcally interacting with thrombin, a crucial serine protease in the blood clotting cascade. As a consequence, cyclotheonamide and chemically synthesized analogs are potentially important for modulating blood clotting. In principle, there is no limitation to the nature of the biological process or cascade, which can be inﬂuenced by (cyclic) peptides and their derivatives, and it should be emphasized that cyclic peptides and mimics derived thereof could be powerful instruments in modulating signal transduction. This is apparent, for example, from the interaction of cyclic peptides with SH2-domains involved in, among others, allergy, cancer and other diseases. At present many biologically extremely important cyclic peptides and derivatives such as octreotide/octreotate [6] fall outside the scope of this contribution, simply because their exact molecular interaction mechanism together with their biological target are still unknown [7], which certainly poses a challenge for future research directed toward unraveling their biomolecular mechanisms of action. In addition, in many cases the molecular events subsequent to the speciﬁc interaction with a receptor are still unknown, leaving many unanswered questions with respect to the ultimate biological effects. Cyclization of a linear peptide (Figure 1.1) endows the cyclic peptide ﬁrst of all with a considerably reduced ﬂexibility as compared to the parent peptide, which is (generally consisting of up to about ten amino acid residues) a very ﬂexible molecule. Each amino acid residue contributes two single bonds to the conformational ﬂexibility of the molecule, and generally speaking the only relatively rigid bond (trans and trans/cis in proline and N-alkylated amino acid-containing peptides) is the amide bond. After approximately ten to twenty amino acid residues, secondary structure elements (such as a-helices [8], turns, and b-strands) start to form. The reduced

Figure 1.1 Different ways of peptide cyclization.

1.1 Introduction

j3

4

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics ﬂexibility of the cyclic peptide derivative as compared to the open form is a distinct advantage for interaction with a potential molecular target. In general, provided that no signiﬁcant enthalpy-entropy compensation takes place [9], a cyclized peptide which is less ﬂexible and therefore more pre-organized, will display a higher afﬁnity because it will lose less entropy upon interaction with its molecular target. As such, cyclization is also a universal ﬁrst approach to increase the afﬁnity of a peptide. An associated advantage of a cyclic peptide structure is the decreased sensitivity to proteolytic degradation, especially by exoproteases, which will be favorable for the half-life of the (cyclic) peptide and thereby its bioavailability.

1.2 Selected Cyclic Peptides

For this review we have selected the cyclic modiﬁed peptides mentioned above, i.e. vancomycin, nisin, cyclosporin, and cyclotheonamide, as well as two examples, i.e. RGD-containing cyclic peptides and SH2 domain binding cyclic peptides, which were inspired by proteins. 1.2.1 Vancomycin

Vancomycin (1, Figures 1.2 and 1.3) is a glycopeptide antibiotic with high activity against Gram-positive bacteria and is particularly renowned for its activity against the feared methicillin-resistant Staphylococcus aureus (MRSA) species [1b]. Vancomycin is produced by Amycolatopsis orientalis, a bacterium originally found in a soil

Figure 1.2 Structure of vancomycin (1). The dipeptide ligand DAla-D-Ala is shown in blue and the hydrogen bonds are shown in red. In the case of the D-Ala-D-lactate ligand, one hydrogen bond is lost and replaced by a repulsion due to the free electron pairs on oxygen.

1.2 Selected Cyclic Peptides

Figure 1.3 Chemical structures of glycopeptides antibiotics vancomycin (1) [1], eremomycin (2) [15a], ristocetin A (3) [15b], and teicoplanin (4) [15c].

sample from Borneo, Indonesia. The biosynthetic pathway and the total syntheses of vancomycin and related structures have been reviewed in a comprehensive account by Nicolaou and coworkers [1b]. Vancomycin is one of the representatives of the glycopeptide family [1]. The vancomycin aglycon consists of a heptapeptide with several non-proteinogenic amino acids. Residues 4, 5, and 7 are phenylglycine derivatives with different substitution patterns of the aromatic ring, while residues 2 and 6 are b-hydroxytyrosines with chlorine substituents in the ortho position. The side chains are involved in the formation of three macrocycles named after the component residues: residues 5 and 7 form the A-B biaryl system, while residues 2, 4, and 6 form the two bisaryl ether systems C-D and D-E, respectively (Figures 1.2 and 1.3). The sugar moieties are glucose and vancosamine [1b]. Vancomycin and other glycopeptides inhibit cell wall synthesis by non-covalent binding to the D-Ala-D-Ala peptide motif of the cell wall precursor lipid II [10]. Four hydrogen bonds are formed when vancomycin binds the D-Ala-D-Ala motif, and all of these involve the peptide backbone [1,10], as shown in Figure 1.2. Dimerization of vancomycin and other glycopeptide antibiotics, as well as anchoring of glycopeptides in the bacterial cell membrane, contribute favorably to the

j5

6

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics antibacterial effect [1]. Although eremomycin (2, Figure 1.3) has a lower afﬁnity for model peptides in vitro, in vivo it is consistently more active than vancomycin. It seems that the actual activity of glycopeptide antibiotics depends on a balance between the dimerization constant and the afﬁnity for model peptides in vitro [11]. The dimerization constant increases when ligand is added and vice versa [12], except in the case of ristocetin A [11] (3, Figure 1.3). Anchoring of the antibiotic to the bacterial cell wall by a lipophilic chain, as is the case with teicoplanin (4, Figure 1.3), also enhances antibacterial activity [13]. In both cases the enhanced activity can be explained by a decrease in entropy of binding, because the binding of the antibiotic to the target peptide becomes effectively intramolecular [1,12,13]. Cell wall synthesis may be inhibited because the enzymes involved in this can no longer bind to the peptide [10]. Resistance to vancomycin can follow three patterns [14]. In both VanA- and VanBtype resistance, the terminal D-alanine residue of the peptidoglycan cell wall precursor is replaced by a D-lactate. Because this changes the amide bond into an ester, a hydrogen bond donor is lost and repulsion between the carbonyl oxygen and lactateester oxygen is present, resulting in a thousand-fold loss of afﬁnity (Figure 1.2). The VanA Enterococci (E. faecium, E. faecalis) are resistant to high concentrations of vancomycin and also clinically used teicoplanin (4), whereas VanB Enterococci are resistant to a wider range of vancomycin concentrations but remain susceptible to teicoplanin [14]. In VanC-type resistance (E. gallinarum, E. casseliﬂavus), the terminal amino acid D-alanine is replaced by D-serine, and loss of activity is probably due to steric hindrance from the serine side chain [1,14]. Knowing the molecular nature of resistance may be very important for the development of new antibiotics to ﬁght resistant micro-organisms. Although the total syntheses of vancomycin and derivatives can be heralded as top achievements in organic synthesis [1], they are of course not practical for obtaining these compounds for application purposes. Therefore, there has been considerable interest in the preparation of simpliﬁed mimics of vancomycin, having especially (part of) the binding cavity for D-Ala-D-Ala. With respect to this, Ellman and coworkers [16], Zhu and coworkers [17], Arnusch and Pieters [18], and Liskamp and coworkers [19] have prepared (monocyclic) mimics of the D-E part of the cavity of these antibiotics via an intramolecular nucleophilic aromatic substitution [16–18] or a Sonogashira-based macrocyclization [19] (Figure 1.4). Recently, a bicyclic mimic of the C-D-E cavity, which was prepared by a Stille reaction followed by tandem ring-closing metathesis (9, Figure 1.4), was described by Liskamp and coworkers [20]. Considerable challenges lie ahead for the synthetic chemist in order to develop practical syntheses of mimics of vancomycin capable of binding not only D-Ala-D-Ala, but also cell wall parts of resistant bacteria, i.e. D-Ala-D-lactate. 1.2.2 Lantibiotic: Nisin

Nisin is a peptide-based antibiotic from the lantibiotic (lanthionine-containing antibiotic) family [2]. It is produced by certain strains of Lactococcus lactis and is active

Figure 1.4 Structures of vancomycin mimics as described by Ellman and coworkers (5) [16], Zhu and coworkers (6) [17], Arnusch and Pieters (8) [18], and Liskamp and coworkers (7) [19] and (9) [20].

1.2 Selected Cyclic Peptides

j7

8

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

Figure 1.5 Representation of the structure of nisin. Hydrophobic amino acid residues are shown in green, polar residues in blue, and unsaturated amino acids in purple. The lanthionine ring is shown in red and the 3-methyllanthionine rings in turquoise.

The structural formulas of lanthionine, 3-methyllanthionine, dehydroalanine, and (Z)-dehydrobutyrine are also shown. The chemical structure of lipid II is shown, and the binding sites of vancomycin and nisin are indicated.

against most Gram-positive bacteria. Its antimicrobial effect has been long known, and nisin has been used as a food preservative for over 30 years. Its structure [21] is characterized by several connected cyclic peptides, including an annulated cyclic peptide system, providing the peptide with a unique shape (or shapes) which could not have been achieved by the corresponding linear sequences (Figure 1.5). Like other antimicrobial peptides, e.g., magainin, nisin has a net positive charge and an amphiphatic character. In addition to the cyclic peptide structures that are so characteristic for the lantibiotics, it contains the unusual amino acids lanthionine and 3-methyllanthionine as well as the unsaturated amino acids dehydroalanine and (Z)-dehydrobutyrine (Figure 1.5). In 1960, Ramseier [22] discovered that nisin causes leakage of intracellular molecules from cells. Later, it was shown that it disturbs the membrane potential and interferes with energy transduction [23]. In addition, it causes inhibition of biosynthesis of the cell wall processes by blocking the synthesis of peptidoglycans [24] and by binding to the precursor lipid II [25]. However, micromolar amounts of nisin are needed to permeate artiﬁcial membranes [26,27] or to inhibit cell wall synthesis in vitro [28], while the in vivo activity of nisin is in the nanomolar range. As vancomycin, which binds to the peptide motif in lipid II, inhibited the antibacterial activity and

1.2 Selected Cyclic Peptides

Figure 1.6 The pyrophosphate cage structure of nisin bound to lipid II. Side chains of nisin are depicted in green, backbone in yellow. Lipid II is depicted in blue [PDB entry code: 1WCO [32], molecular graphics created with YASARA (www.yasara.org) and PovRay (www.povray.org)].

membrane leakage by nisin in intact cells, Breukink et al. [26] concluded that lipid II is necessary for a high speciﬁc nisin activity and resulting pore formation. Studies with artiﬁcial membranes containing lipid II conﬁrmed this conclusion [26]. The N-terminal fragment of nisin (1–12) has been shown to act as a nisin antagonist [29]. Mutation data have shown that changes in the N-terminal fragment reduced nisin activity [30]. NMR data showed that solvent accessibility of the A, B, and C rings of nisin decreased when it bound lipid II [31]. These data independently led to the conclusion that the N-terminal fragment of nisin, notably the ring structure A to C, is involved in lipid II binding. This was conﬁrmed when the NMR structure of the nisin-lipid II complex was determined [32]. Rings A and B of nisin form a cagelike structure around the pyrophosphate moiety of lipid II. This structure shows that hydrogen bonds are formed between backbone amides of nisin and phosphate oxygens, while side chain interactions are of only minor importance. Only leucine residue six is conserved as a hydrophobic residue and interacts with the prenyl chain of lipid II (Figure 1.6). In a model for pore formation, Hsu et al. proposed that after initial binding of nisin to the lipid II pyrophosphate, the C-ring embeds in the membrane, followed by a turn around the ﬂexible hinge region and insertion of the C-terminus [31]. The size of these pores was calculated to be about 2 nm [33]. Studies with pyrene-labeled lipid II enabled Hasper et al. to calculate that a single nisin-lipid II pore consists of four lipid II molecules and eight nisin molecules (Figure 1.7). The pores are highly stable, even on addition of detergents that cause disruption of membranes [34]. Other lantibiotics that do not have the ability to form pores and indeed do not cause cell leakage have nevertheless still an impressive antibacterial effect in vivo. With respect to this, an alternative mechanism of antibacterial activity was recently described by which members of the lantibiotic family kill Gram-positive bacteria by binding lipid II and removing it from the cell division site (or septum) and thus block cell wall synthesis [35]. The cyclic peptide systems in nisin can be considered as side chain-cyclized peptide derivatives. However, they do not contain disulﬁde bridges, which is the most common

j9

10

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

Figure 1.7 Interaction of nisin with phospholipid model membranes: mechanism of pore formation (reprinted with permission from Ref. [34]. Copyright 2007 American Chemical Society).

way of generating side chain-cyclized peptides. In contrast they have sulﬁde or thioether bridges as part of their ring structures, which are introduced biosynthetically by posttranslational modiﬁcation. This rather unusual cyclization moiety enticed us to start investigating whether also bridges other than thioether bridges could be introduced as cyclization elements for obtaining mimics of the nisin ring structures [36]. Attention was directed towards designing and synthesizing alkene, alkyne, and alkane mimics of the thioether bridge, using ring-closing alkene and alkyne metathesis, sometimes followed by hydrogenation. Thus mimics of the A (10, 11, and 14), AB (12), C, and D/E (16) ring systems were prepared (Scheme 1.1). Apparently, the backbone structure of the D/E ring system has a predisposition towards a knotted ring structure, since it was possible to obtain an alkene mimic (16) directly in good yield (95%) from a double ring closing metathesis reaction of the tetra-allylglycine containing linear precursor (15) in a single reaction step [36d] (Scheme 1.1). Based on the pyrophosphate cage structure of nisin bound to lipid II, a tricyclic mimic was designed and synthesized in which a lactam bridge connects the B-ring mimic with the A-ring mimic. The biological activity data are very promising and justify further improvement of the designed lipid II pyrophosphate binders. The inherent ﬂexibility of a peptide, even when it is cyclized, challenges us to control its shape by alternative constraints to those already present. 1.2.3 Cyclosporin A

Cyclosporin was isolated from the fungus Tolypocladium inﬂatum. Its immunosuppressive activity was discovered in 1976 [4]. Although since then many analogs have been prepared and investigated, cyclosporin A (CsA, Sandimmune, 17) remains the most effective cyclic peptide and is the major immunosuppressant drug to prevent graft rejection after transplant surgery. CsA is a cyclic undecapeptide consisting completely of hydrophobic amino acids, as shown in Figure 1.8 [37]. Additional structural features are a threonine-derived butenyl-containing amino acid derivative as well as six N-methylated amino acid residues. In addition to reducing the proteolytic degradation rate and increasing the

1.2 Selected Cyclic Peptides

Scheme 1.1 Synthesis and structures of alkane- and alkyne-bridged nisin mimics [36].

Figure 1.8 Structural formulas of cyclosporin A and FK506.

j11

12

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics hydrophobicity, the inﬂuence of these N-methylated residues on the ﬂexibility of the cyclic peptide system could be interesting. The molecular basis of CsAs biological activity is the formation of a complex with the immunophilin cyclophilin A (CyPA) in such a way that the CsA-CyPA complex can associate with calcineurin (CN), which is a Ca2þ/calmodulin-dependent Ser/Thr phosphatase. As a result, dephosphorylation of a nuclear factor of activated T-cells (NFAT) is inhibited. X-ray structures [38] of the CsA-CyPA complex show that one side of the CsA ring, comprising the residues 1–3 and 9–11, ﬁts into a groove on CyPA, leaving the opposite side free to interact with CN. Interestingly, CyPA is a cis-trans proline isomerase, but this enzymatic activity does not play a role in CN inhibition. It seems that CsA interacts with CyPA to create a larger complex capable of interacting with CN. X-ray structures of the CsA-CyPA-CN complexes (Figure 1.9) show that the conformation of the CsA backbone and the side chains interacting with CyPA in this ternary complex are very similar to those in the binary CsA-CyPA complex. However, the side chains of the CNbinding moiety of CsA have undergone signiﬁcant conformational changes. CsA as well as FK506 (18), a completely unrelated non-peptide natural product with similar activity to CsA, have resisted many efforts to improve them as drugs. However, the X-ray structures of the ternary complex may help to stimulate the design of superior immunosuppressants.

Figure 1.9 Ribbon representation of CyPA-CsA-CN. The catalytic subunit (CNA) of CN is shown in gold, the regulatory subunit (CNB) of CN cyan, cyclosporin (CsA) green, cyclophylin (CyPA) red, the Zn2þ and Fe3þ ions pink, and the Ca2þ ions blue. The residues from CN involved in binding of CyPA-CsA are shown as a blue ball-stick representation (reprinted with permission from Ref. [38]. Copyright 2007 National Academy of Sciences, USA.)

1.2 Selected Cyclic Peptides

1.2.4 Cyclotheonamide A and B

In the early 1990s, Fusetani and Matsunaga isolated two substances from marine sponges of the genus Theonella, which were called cyclotheonamide A and B (CtA 19 and CtB 20, respectively), and were found to be potent inhibitors of various proteinases, especially the blood coagulation enzyme thrombin [5,39]. After structural elucidation by extensive NMR analysis [40], the cyclotheonamides were found to be cyclic pentapeptides having a 19-membered ring and containing a D-phenylalanine (D-Phe), an L-proline (L-Pro), an L-2,3-diaminopropionic acid (L-Dpr) with either a formylated (For, CtA 19) or an acetylated (Ac, CtB 20) N2-amino moiety, a vinylogous L-tyrosine (L-vTyr), and an a-ketoarginine (L-kArg) residue (Figure 1.10). Noteworthy are the presence of the uncommon non-proteinogenic amino acids L-a-ketoarginine (a-oxo-L-b-homoarginine) and vinylogous L-tyrosine (L-tyrosinederived a,b-unsaturated g-amino acid). More important, however, is the presence of the a-ketoamide moiety indicating the possible mechanism of action of the cyclotheonamides as potent serine protease inhibitors by serving as an electrophilic trap to form a tetrahedral intermediate (a transition state analog) with the hydroxyl group of Ser195, part of the catalytic triad His57-Asp102-Ser195 of the proteinase. To obtain insight into the molecular basis of action of the cyclotheonamides and to explain the high afﬁnity as a serine protease inhibitor, complexes of CtA with a-thrombin [41a,c] and b-trypsin [41b,c], respectively, were studied by X-ray crystallography. Thrombin is the pivotal trypsin-like protease for the regulation of thrombosis and hemostasis. Thrombin hydrolyzes its natural substrates by recognition of the Pro-Arg motif in the apolar S2- and the primary speciﬁcity S1 pocket [42]. The molecular structure of the thrombin-CtA complex (Figure 1.11) showed that CtA was bound to the active site of the enzyme. The arginine side chain formed an electrostatic interaction with Asp189, located at the bottom of the S1 binding pocket.

Figure 1.10 Structure of cyclotheonamide A and B.

j13

14

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

Figure 1.11 Stereoview of CtA 19 bound in the active-site cleft of a-thrombin [PDB entry code: 1TMB [41a], molecular graphics created with YASARA (www.yasara.org) and PovRay (www. povray.org)].

Furthermore, the backbone of the Pro-Arg segment formed a hydrogen-bonded antiparallel b-chain with the Ser214-Gly216 sequence in thrombin. Both features were highly similar to what was observed for the binding of the highafﬁnity active site-directed thrombin inhibitor H-D-Phe-Pro-Arg chloromethyl ketone [42]. A well-deﬁned continuous electron density between Ser195 and the carbonyl carbon of the a-ketoamide was observed, which was a strong indication of a covalent complex through the formation of a tetrahedral hemiketal. The other carbonyl oxygen was involved in a hydrogen-bonding network in the oxyanion hole formed by Gly193 and Ser195. X-ray analysis of the trypsin-CtA complex showed that the gross structural features were similar to thrombin. However, the higher degree of speciﬁcity for trypsin (Ki 0.023 mM versus 0.18 mM) is fairly well explained by the Trp60A-Thr60I insertion loop of thrombin [42], which narrows the active-site binding cleft and thus results in less favorable interaction with the D-Phe and L-vTyr residues. The interesting macrocyclic pentapeptide scaffold prompted several research groups to design new total synthesis routes for cyclotheonamide [40,41a,b,43] and derivatives thereof [44], mainly for SAR-studies for targeting the development of novel synthetic thrombin inhibitors. The ﬁrst total synthesis of CtB 20 was published by Hagihara and Schreiber [40]. This synthesis (Scheme 1.2) deﬁned the true stereochemical structure of cyclotheonamide, which had previously been reported incorrectly. The linear CtB precursor 21 was synthesized from dipeptide Ac-Dpr(Boc)-ProOPac 22, following three amide coupling reactions in the C ! N direction. The vinylogous tyrosine derivative 23 was synthesized from the protected Weinreb amide 24 that was obtained from the aldehyde intermediate in a Wittig homologation followed by conversion into an active pentaﬂuorophenyl ester. Boc-D-Phe-OH was coupled with BOP/DMAP as the condensing agents. The a-hydroxy acid 25 was prepared from the protected arginine Weinreb amide derivative 26. To avoid an intramolecular nucleophilic attack of the guanidinium moiety at the intermediate arginine aldehyde, it was doubly protected with Boc and Mtr groups. Finally, macrolactamization was achieved by a four-step process. Deprotection of the

1.2 Selected Cyclic Peptides

Scheme 1.2 Retrosynthesis of the ﬁrst total synthesis of CtB 20 as described by Hagihara and Schreiber [40].

phenacyl ester was followed by conversion into the active pentaﬂuorophenyl ester. Then, the N-terminal Boc functionality was selectively removed by pTsOH. Neutralization with DIPEA and treatment with DMAP resulted in a smooth conversion into the protected peptide macrocycle. After oxidation of the hydroxyl functionality, all remaining protecting groups were removed by treatment with TFA in the presence of thioanisole as scavenger to give CtB 20 in 3% overall yield over 16 reaction steps. The total synthesis of CtA 19, which was used in the ﬁrst X-ray studies to elucidate the structure of the CtA – a-thrombin complex (Figure 1.11), is shown in Scheme 1.3. This synthesis, as described by Maryanoff et al. [41a], was based on a convergent [3 þ 2] fragment condensation route (Scheme 1.3). The two segments 28 and 29 were synthesized from their requisite protected amino acid building blocks 30, 31 and 32, 33 respectively. Coupling of 28 and 29 to give linear intermediate 27 proceeded in 65% yield. After selective removal of the N- and C-terminal protecting groups on 27, the critical macrocyclization was effected in 65% yield under high dilution conditions. The intermediate protected macrocycle was then processed in four steps (removal of the phthaloyl functionality and subsequent formylation of the a-amino group, Dess-Martin oxidation to the a-ketoamide, and ﬁnally treatment with HF to remove the tosyl protecting group) to give CtA 19. Since also many linear peptides that are powerful inhibitors of thrombin [45], have been synthesized it is difﬁcult to evaluate to what extent the cyclic nature of peptide in cyclotheonamide contributes to the activity. Nevertheless, the cyclic nature is advantageous for possible in vivo activity, because of the absence of polar end groups and the greater steric hindrance. With respect to the latter, this can confer greater resistance to degradation by peptidases and enhanced ability to traverse membrane barriers. Unfortunately, cyclotheonamide and congeners have not been further developed into suitable compounds for anticoagulation therapy.

j15

16

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

Scheme 1.3 Retrosynthesis of the total synthesis of CtA 19, which was used in the X-ray analysis as a complex with a-thrombin [41a].

1.2.5 cyclo RGD Peptides as aVb3 Antagonists

So far, naturally occurring cyclic peptides and mimics thereof have been discussed, in which the cyclic character plays an essential role with respect to afﬁnity and activity in particular. However, this could also be the case for suitable sequences present in proteins, which upon incorporation into macrocyclic peptide structures could lead to molecular constructs with similar or even enhanced afﬁnity and activity without the requirement for the entire protein or a large derivative. This is an exceptional challenge, which could lead to a variety of new biologically active peptides. Although it is a realistic strategy, as is evidenced from the vast number of biologically active cyclic peptides, it may be sometimes a prolonged process and an integral approach featuring an iterative design-synthesis-activity-structure cycle seems to be essential. Two important examples of cyclic peptides will be highlighted in which amino acid sequences of larger proteins have been successfully used as part of macrocyclic peptide structures to mimic in an almost pars pro toto approach important properties of the parent protein.

1.2 Selected Cyclic Peptides

Integrins are a class of heterodimeric transmembrane proteins [46] which play an important role in cell signaling, cell-cell adhesion, apoptosis, and cell-matrix interactions [47]. Integrin aVb3, which binds to Arg-Gly-Asp (RGD) tripeptide motif-containing ligands [48], plays a pivotal role in tumor angiogenesis and metastasis and is expressed on activated endothelial cells during tumor-induced angiogenesis and on various tumor cell types (e.g., breast, ovarian, and prostate cancers), whereas it is absent on quiescent endothelial cells and normal tissues [47]. Evidence exists that inhibition of aVb3 integrin function prevents tumor growth and induces tumor regression by antagonizing angiogenesis [49]. Several peptidic [3] and peptidomimetic [50] aVb3 antagonists have been synthesized. Among these, the cyclo[Arg-Gly-Asp-D-Phe-Val] (c[RGDfV]) (34), as developed by Kessler and coworkers, is one of the most active and selective antagonists for the aVb3 integrin [51] (Figure 1.12). Structure-activity relationship studies on this cyclic pentapeptide showed that the exchange of the valine by a lysine residue (Lys, K) (compound 36) did not signiﬁcantly inﬂuence activity and selectivity [52]. Because the e-amino moiety of the lysine residue can easily be modiﬁed, numerous applications of c[RGDfK] have been studied for tumor targeting and imaging [53]. The linear precursor of, for example, 34 [51] was obtained by solid-phase peptide synthesis performed on an SASRIN -resin applying the Fmoc/tBu strategy. This allowed the preparation of a protected peptide with both free amino and carboxylic acid termini, which were head-to-tail cyclized using diphenylphosphorylazide, leading to the protected cyclic peptide. Removal of side chain protecting groups afforded 34. The structural basis for binding of the prototypical Arg-Gly-Asp sequence to the aVb3 integrin was recently provided by the crystal structure of 35 complexed with the extracellular segment of integrin aVb3 [48] as shown in Figure 1.13.

Figure 1.12 Structures of cyclic pentapeptides as potent antagonists of the aVb3 integrin.

j17

18

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

Figure 1.13 The ligand-integrin binding site. (A) Surface representation of the ligandbinding site, with the ligand peptide (35, Fig. 1.12) shown as ball-and-stick model. (B) Interactions between ligand and integrin. The peptide (yellow) and residues interacting with the ligand or with Mn2þ ions are shown in ball-and-stick representation. aV and b3

residues are labeled blue and red, respectively. Oxygen and nitrogen atoms are in red and blue, respectively. Hydrogen bonds and ˚ ) are resalt bridges (distance cutoff, 3.5 A presented with dotted lines (reprinted with permission from Ref. [48]. Copyright 2007 American Association for the Advancement of Science).

The interaction with the Arg-Gly-Asp ligand triggers the structural rearrangements of the aVb3 integrin which are necessary for cell signaling pathways. Cyclic RGD peptide 35 binds at the major interface between the aV and b3 subunits. The arginine side chain inserts into a narrow groove forming a bidentate salt bridge to Asp218 and Asp150. The carboxylate moiety and the aspartic acid residue of 35 form an extensive network of polar interactions with the protein and are completely buried in the complex. The glycine residue undergoes several hydrophobic interactions and lies at the interface between the a and b subunits. The other two residues of the cyclic RGD pentapeptide point into the solution and do not have structural interactions with the aVb3 integrin. Comparison of the X-ray structures of the unbound

1.2 Selected Cyclic Peptides

and the RGD-bound aVb3 integrin revealed many conformational rearrangements within the protein which may form a structural basis for receptor activation. The backbone conformation of the Arg-Gly-Asp tripeptide motif in the pentapeptide is almost identical to that of the RGD sequence in the natural ligand Echistatin, which is a 49 amino acid protein isolated from the venom of Echis carinatus and is one of the smallest natural ligands that interacts with integrin-type receptors through an Arg-Gly-Asp sequence [54]. This suggests that this X-ray structure [48] (Figure 1.13) may provide a basis for obtaining insight into the molecular interaction details of integrins with other and larger RGD-containing ligands. 1.2.6 SH2 Domain-Binding Peptides

The class of Src homology 2 (SH2) domains is the largest class of phosphotyrosinerecognizing protein domains. They occur in different protein groups that are mostly involved in signaling and regulation, and thus perform several functions within the cell [55]. Mutations in SH2 domains are implicated in many human diseases, of which a complete overview is given by Liu et al. [55a]. The SH2 domains are approximately 100 amino acids in size and have a well-deﬁned structure, consisting of a central antiparallel b-sheet and two ﬂanking a-helices, followed by a second, two-strand sheet, as elucidated by X-ray crystallography and NMR spectroscopy (Figure 1.14) [56,57].

Figure 1.14 NMR-structure of the Grb2-SH2 domain in complex with inhibitor 39. A central antiparallel b-sheet in the protein is ﬂanked by two a-helices, with peptide binding mediated by the sheet, intervening loops, and one of the helices. The speciﬁc recognition of phosphotyrosine involves cation-p interactions between

lysine and arginine side chains and the aromatic tyrosine side chain, in addition to hydrogen-bond interactions with the phosphate moiety [57] [PDB entry code: 1X0N [57], molecular graphics created with YASARA (www.yasara.org) and PovRay (www.povray. org)].

j19

Figure 1.15 Structures of synthetic Grb2 SH2 domain-binding analogs [58—60].

20

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics

1.3 Selected Cyclic Peptides

The recognition of phosphorylated peptides by SH2 domains depends on the phosphorylated residue. Speciﬁcity is determined by one or two nearby residues. Apart from the phosphotyrosine-speciﬁc recognition, SH2-domains can interact in different ways with other proteins [55]. This paragraph focuses on the interactions mediated by the conventional phosphopeptide recognition region. (The Grb2 SH2 domain plays a crucial role in the Ras signal transduction pathway, which is an important regulator of cell growth and differentiation.) Although most SH2 domains bind their phosphorylated peptide ligands in an extended, b-strandlike conformation, the Grb2 SH2 domain binds the natural ligand in a type I bturn conformation [58]. High-afﬁnity binding and recognition processes are mediated by the pTyr-Xaa-Asn motif. In tripeptide 37 (Figure 1.15), the amino acid at the second position is a 1-aminocyclohexanecarboxylic acid residue, which induces a b-turn conformation. This results in a high binding afﬁnity of the peptide for the Grb2 SH2 domain. Removal of the acetylated amino functionality of the pTyr residue (compound 38, Figure 1.15) led to a 100-fold lower afﬁnity, but cyclization restored this loss. However, in an intracellular assay, the cyclic peptide without a substitution (compound 39, Figure 1.15) was markedly less active than the linear peptide with this substitution [58]. A KD value of 75 pM was reported when the naphthyl group was replaced by a 5-methylindole and the a substitution was restored (compound 41, Figure 1.15). In cell assays, the compound was an effective inhibitor of mitosis in breast cancer cell culture [59]. Cyclization of peptides containing phosphate group mimics (42) increased binding afﬁnities about ﬁve-fold in ELISA assays compared to their linear equivalents [60]. However, one should not consider cyclization as a panacea. Recently, we found, for example, that cyclization by ring-closing metathesis of a phosphopeptide interacting with the Grb2 SH2 domain did not give rise to improvement of binding afﬁnity (Figure 1.16) [61].

Figure 1.16 Linear phosphopeptide 43 and cyclic phosphopeptide ligands 44 and 45 obtained by ring-closing metathesis showing virtually no difference in afﬁnity for the Grb2 SH2 domain as measured by a surface plasmon resonance assay [61].

j21

22

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics 1.3 Conclusions

The very limited number of bioactive (macro)cyclic peptides and peptide mimics for which molecular details of the interactions with targets are known stands in striking contrast to the vast and still increasing number of bioactive cyclic peptides. This calls for approaches to the elucidation of molecular mechanisms of interaction and target identiﬁcation, since the biological activity of many of those peptides is extremely interesting, varying from antibiotic or anti-cancer activity to epigenetic behavior, or activity in signal transduction mechanisms. The studies could form a basis for the design and development of new generations of bioactive compounds also capable of interacting with large surfaces involved in many important protein-protein interactions. In this respect, macrocyclic peptide structures are particularly interesting because, in addition to the possibility of offering a large surface, the amino acid residues of which they are composed can offer a large diversity of functionalities for interaction. Since, in principle, there are no limitations to the preparation of macrocyclic peptide molecular constructs, there are plenty of opportunities for design and synthesis in the promising bio-supramolecular chemistry area.

1.4 Experimental: Selected Procedures 1.4.1 Synthesis of Bicyclic Peptide 9: an Alkene-bridged Mimic of the Vancomycin C-D-E Cavity

The original synthesis of this compound was described in Ref. [20]. The linear pentapeptide precursor, Na-(tert-butyloxycarbonyl)-seryl(allyl)-D-alanyl-R,S-4-hydroxy-3,5bisallyl-phenylglycyl-D-leucyl-serine(allyl) methyl amide (20 mg, 0.025 mmol) was dissolved in a mixture of TCE (12 mL) and DMF (0.5 mL). This solution was purged with N2 and heated to 80 C followed by the addition of the 2nd generation Grubbs catalyst (21 mg, 0.025 mmol). After 10 min the reaction was complete according to TLC and ES-MS. Subsequently, the solvent was removed under reduced pressure and the residue was puriﬁed by column chromatography (eluent: CH2Cl2 ! CH2Cl2/ MeOH 95:5, v/v). The tandem ring-closed pentapeptide 9 was obtained as a slightly brownish solid in 67% yield (13 mg). HPLC analysis showed that the product was 92% pure (based on the diastereomers). HR-MS calcd. for [M þ H]þ: C37H54N6Oþ 10 : 743.3979; found m/z [M þ H]þ: 743.3969. 1.4.2 Synthesis of Cyclic Peptide 14: an Alkyne-bridged Mimic of the Nisin A-Ring Fragment

The original synthesis of this compound was described in Ref. [36b]. The ringclosing alkyne metathesis reaction was carried out under argon in ﬂame-dried

Abbreviations

glassware using Schlenk techniques. The linear pentapeptide precursor, Na-(tertbutyloxycarbonyl)-(2-butyne-glycyl)-isoleucyl-alanyl-leucyl-(2-butyne-glycine) methyl ester (46 mg, 0.070 mmol) was dissolved in dry toluene (200 mL), and the tungstenalkylidyne catalyst (tBuO)3W CtBu (4 mg, 9 mmol) was added. The obtained reaction mixture was heated to 80 C and stirred for 2 h. The reaction was monitored by TLC until no changes in product distribution could be observed. Then, H2O (1 mL) was added to quench the catalyst, and the solvent was removed by evaporation. The residue was puriﬁed by column chromatography with CH2Cl2/MeOH 97.5:2.5, v/v as the eluent. Cyclic pentapeptide 14 was obtained as an off-white powder in 42% yield (18 mg). HR-MS calcd for [M þ H]þ: C29H47N5Oþ 8 : 616.33218; found m/z [M þ H]þ: 616.33223.

Abbreviations

Abu Ac Ala Arg kArg Asn Asp Boc BOP Bn Bu t Bu Cbz CN CsA CtA CtB CyPA Dha Dhb DIPEA DMAP DMF Dpr ELISA ES-MS Fmoc For GlcNAc i

2-aminobutyric acid acetyl alanine (A) arginine (R) a-ketoarginine; a-oxo-L-b-homoarginine asparagine (N) aspartic acid (D) tert-butyloxycarbonyl benzotriazol-1-yloxy-tris(dimethylamino)phosphonium hexaﬂuorophosphate benzyl iso-butyl tert-butyl carbobenzyloxy calcineurin cyclosporin A cyclotheonamide A cyclotheonamide B cyclophilin A dehydroalanine (Z)-dehydrobutyrine diisopropylethylamine 4-dimethylaminopyridine dimethylformamide 2,3-diaminopropionic acid enzyme linked immunosorbent assay electrospray mass spectrometry 9-ﬂuorenylmethyloxycarbonyl formyl N-acetyl glucosamine

j23

24

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics Gly Grb2 His HPLC HR-MS Ile Leu Lys Me Met MRSA Mtr MurNAc NFAT NMR Pac Phe Pht Pro RCAM RCM SAR SASRIN SEM Ser SH2 TBDMS TCE TFA Thr TLC Tos Trp Trt pTsOH Tyr vTyr Val VanA VanB VanC Xaa

glycine (G) growth factor receptor-bound protein 2 histidine (H) high-pressure liquid chromatography high-resolution mass spectrometry isoleucine (I) leucine (L) lysine (K) methyl methionine (M) methicillin resistant Staphylococcus aureus 4-methoxy-2,3,6-trimethylbenzenesulfonyl N-acetyl muramic acid nuclear factor of activated T-cells nuclear magnetic resonance phenacyl phenylalanine (F) phthaloyl proline (P) ring-closing alkyne metathesis ring-closing metathesis structure-activity relationship 2-amino-4-alkoxybenzylalcohol resin 2-(trimethylsilyl)ethoxymethyl serine (S) Src homology 2 tert-butyl dimethyl silyl 1,1,2-trichloroethane triﬂuoroacetic acid threonine (T) thin layer chromatography tosyl tryptophan (W) trityl para-toluenesulfonic acid tyrosine (Y) vinylogous tyrosine; L-tyrosine-derived a,b-unsaturated g-amino acid valine (V) resistance to vancomycin of type A resistance to vancomycin of type B resistance to vancomycin of type C unspeciﬁed amino acid residue

References

References 1 (a) Boger, D.L. (2001) Med. Res. Rev., 21 (5), 356–381; (b) Nicolaou, K.C., Boddy, C.N.C., Bräse, S. and Winssinger, N. (1999) Angew. Chem., 111, 2230–2287. Angew. Chem. Int. Ed., 38, 2096–2152; (c) Rao, A.V.R., Gurjar, M.K., Reddy, K.L. and Rao, A.S. (1995) Chem. Rev., 95, 2135–2167. 2 Chatterjee, C., Paul, M., Xie, L. and van der Donk, W.A. (2005) Chem. Rev., 105, 633–684. 3 Haubner, R., Finsinger, D. and Kessler, H. (1997) Angew. Chem., 109, 1440–1456; Angew. Chem. Int. Ed., 36, 1374–1389. 4 Rüegger, A., Kuhn, M., Lichti, H., Loosli, H.-R., Huguenin, R., Quiquerez, C. and von Wartburg, A. (1976) Helv. Chim. Acta, 59, 1075–1092. 5 Fusetani, N., Matsunaga, S., Matsumoto, H. and Takebayashi, Y. (1990) J. Am. Chem. Soc., 112, 7053–7054. 6 Froideveaux, S. and Eberle, A.N. (2002) Biopolymers (Peptide Science), 66, 161–183. 7 X-ray structure of octreotide: Pohl, E., Heine, A., Sheldrick, G.M., Dauter, Z., Wilson, K.S., Kallen, J., Huber, W. and Pfäfﬂi, P.J. (1995) Acta Crystallogr., Sect. D, 51, 48–59. 8 Rohl, C.A., Chakrabartty, A. and Baldwin, R.L. (1996) Protein Science, 5, 2623–2637. 9 (a) Benﬁeld, A.P., Teresk, M.G., Plake, H.R., DeLorbe, J.E., Millspaugh, L.E. and Martin, S.F. (2006) Angew. Chem., 118, 6984–6989; Angew. Chem. Int. Ed., 45, 6830–6835; (b) de Mol, N.J., Catalina, M.I., Dekker, F.J., Fischer, M.J.E., Heck, A.J.R. and Liskamp, R.M.J. (2005) ChemBioChem., 6, 2261–2270. 10 Knox, J.R. and Pratt, R.F. (1990) Antimicrob. Agents Chemother., 34, 1342–1347. 11 Mackay, J.P., Gerhard, U., Beauregard, D.A., Maplestone, R.A. and Williams, D. H. (1994) J. Am. Chem. Soc., 116, 4573– 4580.

12 Williams, D.H., Maguire, A.J., Tsuzuki, W. and Westwell, M.S. (1998) Science, 280, 711–714. 13 Mackay, J.P., Gerhard, U., Beauregard, D.A., Westwell, M.S., Searle, M.S. and Williams, D.H. (1994) J. Am. Chem. Soc., 116, 4581–4590. 14 (a) Süssmuth, R.D. (2002) ChemBioChem, 3, 295–298; (b) Walsh, C. T., Fisher, S.L., Park, I.-S., Prahalad, M. and Wu, Z. (1996) Chem. Biol., 3, 21–28. 15 (a) Printsevskaya, S.S., Pavlov, A.Y., Olsufyeva, E.N., Mirchnink, E.P., Isakova, E.B., Reznikova, M.I., Goldman, R.C., Branstrom, A.A., Baizman, E.R., Longley, C.B., Sztaricskai, F., Batta, G. and Preobrazhenskaya, M.N. (2002) J. Med. Chem., 45, 1340–1347; (b) Crowley, B.M., Mori, Y., McComas, C.C., Tang, D. and Boger, D.L. (2004) J. Am. Chem. Soc., 126, 4310–4317; (c) Boger, D.L., Kim, S. H., Mori, Y., Weng, J.-H., Rogel, O., Castle, S.L. and McAtee, J.J. (2001) J. Am. Chem. Soc., 123, 1862–1871. 16 Xu, R., Greiveldinger, G., Marenus, L.E., Cooper, A. and Ellman, J.A. (1999) J. Am. Chem. Soc., 121, 4898–4899. 17 Bois-Choussy, M., Neuville, L., Beugelmans, R. and Zhu, J. (1996) J. Org. Chem., 61, 9309–9322. 18 Arnusch, C.J. and Pieters, R.J. (2003) Eur. J. Org. Chem., 3131–3138. 19 ten Brink, H.T., Rijkers, D.T.S. and Liskamp, R.M.J. (2006) J. Org. Chem., 71, 1817–1824. 20 ten Brink H.T., Rijkers D.T.S., Kemmink J., Hilbers H.W. and Liskamp R.M.J. (2004) Org. Biomol. Chem., 2, 2658–2663. 21 Gross, E. and Morell, J.L. (1971) J. Am. Chem. Soc., 93, 4634–4635. 22 Ramseier, H.R. (1960) Arch. Microbiol., 37, 57–94. 23 Ruhr, E. and Sahl, H.-G. (1985) Antimicrob. Agents Chemother., 27, 841–845.

j25

26

j 1 Bioactive Macrocyclic Peptides and Peptide Mimics 24 Linnett, P.E. and Strominger, J.L. (1973) Antimicrob. Agents Chemother., 4, 231–236. 25 Reisinger, P., Seidel, H., Tschesche, H. and Hammes, W.P. (1980) Arch. Microbiol., 127, 187–193. 26 Breukink, E., Wiedemann, I., van Kraaij, C., Kuipers, O.P., Sahl, H.-G. and de Kruijff, B. (1999) Science, 286, 2361–2364. 27 Hoffmann, A., Pag, U., Wiedemann, I. and Sahl, H.-G. (2001) Il Farmaco, 57, 685–691. 28 Brötz, H., Josten, M., Wiedemann, I., Schneider, U., Götz, F., Bierbaum, G. and Sahl, H.-G. (1998) Mol. Microbiol., 30, 317–327. 29 Chan, W.C., Leyland, M., Clark, J., Dodd, H.M., Lian, L.-Y., Gasson, M.J., Bycroft, B.W. and Roberts, G.C.K. (1996) FEBS Lett., 390, 129–132. 30 Wiedemann, I., Breukink, E., van Kraaij, C., Kuipers, O.P., Bierbaum, G., de Kruijff, B. and Sahl, H.-G. (2001) J. Biol. Chem., 276, 1772–1779. 31 Hsu, S.-T.D., Breukink, E., de Kruijff, B., Kaptein, R., Bonvin, A.M.J.J. and van Nuland, N.A.J. (2002) Biochemistry, 41, 7670–7676. 32 Hsu, S.-T.D., Breukink, E., Tischenko, E., Lutters, M.A.G., de Kruijff, B., Kaptein, R., Bonvin, A.M.J.J. and van Nuland, N.A.J. (2004) Nat. Struct. Mol. Biol., 11, 963–967. 33 van Heusden, H.E., de Kruijff, B. and Breukink, E. (2002) Biochemistry, 41, 12171–12178. 34 Hasper, H.E., de Kruijff, B. and Breukink, E. (2004) Biochemistry, 43, 11567–11575. 35 Hasper, H.E., Kramer, N.E., Smith, J.L., Hillman, J.D., Zachariah, C., Kuipers, O. P., de Kruijff, B. and Breukink, E. (2006) Science, 313, 1636–1637. 36 (a) Ghalit, N., Rijkers, D.T.S., Kemmink, J., Versluis, C. and Liskamp, R.M.J. (2005) Chem. Commun., 192–194; (b) Ghalit, N., Poot, A.J., Fürstner, A. and Rijkers, D.T.S. and Liskamp, R.M.J. (2005) Org. Lett., 7, 2961–2964; (c)

37

38

39

40

41

42

43

Ghalit, N., Rijkers, D.T.S. and Liskamp, R.M.J. (2006) J. Mol. Catal. A: Chem., 254, 68–77; (d) Ghalit, N., Kemmink, J., Hilbers, H.W., Versluis, C., Rijkers, D.T. S. and Liskamp, R.M.J. (2007) Org. Biomol. Chem., 5, 924–934. For a solid phase synthesis: Angell, Y.M., Thomas, T.L., Flentke, G.R. and Rich, D. H. (1995) J. Am. Chem. Soc., 117, 7279–7280. Huai, Q., Kim, H.-Y., Liu, Y., Zhao, Y., Mondragon, A., Liu, J.O. and Ke, H. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 12037–12042. For a review on sponge peptides: Fusetani, N. and Matsunaga, S. (1993) Chem. Rev., 93, 1793–1806. Since the original paper by Fusetani and Matsunaga [39], Hagihara and Schreiber reassigned the stereochemistry of the vinylogous tyrosine- and a-ketoarginine residue to be of the S-conﬁguration Hagihara, M. and Schreiber, S.L. (1992) J. Am. Chem. Soc., 114, 6570–6571. (a) Maryanoff, B.E., Qiu, X., Padmanabhan, K.P., Tulinsky, A., Almond, H.R. Jr., Andrade-Gordon, P., Greco, M.N., Kauffman, J.A., Nicolaou, K. C., Liu, A., Brungs, P.H. and Fusetani, N. (1993) Proc. Natl. Acad. Sci. U.S.A., 90, 8048–8052; (b) Lee, A.Y., Hagihara, M., Karmacharya, R.,Albers, M.W., Schreiber, S.L. and Clardy, J. (1993) J. Am. Chem. Soc., 115, 12619–12620; (c) Ganesh, V., Lee, A. Y., Clardy, J. and Tulinsky, A. (1996) Protein Sci., 5, 825–835. Bode, W., Mayr, I., Baumann, U., Huber, R., Stone, S.R. and Hofsteenge, J. (1989) EMBO J., 8, 3467–3475. (a) Wipf, P. and Kim, H. (1993) J. Org. Chem., 58, 5592–5594; (b) Deng, J., Hamada, Y., Shioiri, T., Matsunaga, S. and Fusetani, N. (1994) Angew. Chem., 106, 1811–1813; Angew. Chem. Int. Ed. Engl., 33, 1729–1731; (c) Maryanoff, B.E., Greco, M.N., Zhang, H.-C., AndradeGordon, P., Kauffman, J.A., Nicolaou, K. C., Liu, A. and Brungs, P.H. (1995) J. Am. Chem. Soc., 117, 1225–1239; (d)

References

44

45

46

47

48

49

50

51

52

53

Bastiaans, H.H.M., van der Baan, J.L. and Ottenheijm, H.C.J. (1997) J. Org. Chem., 62, 3880–3889. (a) Maryanoff, B.E., Zhang, H.-C., Greco, M.N., Glover, K.A., Kauffman, J.A. and Andrade-Gordon, P. (1995) Bioorg. Med. Chem., 3, 1025–1038; (b) Nakao, Y., Matsunaga, S. and Fusetani, N. (1995) Bioorg. Med. Chem., 3, 1115–1122. For a review see: Gustafsson, D., Bylund, R., Antonsson, T., Nilsson, I., Nyström, J.-E., Eriksson, U., Bredberg, U. and Teger-Nilsson, A.-C. (2004) Nat. Rev. Drug Discovery, 3, 649–659. (a) Plow, E.F., Haas, T.A., Zhang, L., Loftus, J. Smith, J.W. (2000) J. Chem. Biol., 275, 21785–21788; (b) Gottschalk, K.-E. and Kessler, H. (2002) Angew. Chem., 114, 3919–3927; Angew. Chem. Int. Ed., 41, 3767–3774. (a) Hynes, R.O. (1992) Cell, 69, 11–25; (b) Brooks, P.C., Clark, R.A. and Cheresh, D.A. (1994) Science, 264, 569–571; (c) Hynes, R.O. (2002) Nat. Med., 8, 918–921. Xiong, J.-P., Stehle, T., Zhang, R., Joachimiak, A., Frech, M., Goodman, S. L. and Arnaout, M.A. (2002) Science, 296, 151–155. Brooks, P.C., Montgomery, A.M.P., Rosenfeld, M., Reisfeld, R.A., Hu, T., Klier, G. and Cheresh, D.A. (1994) Cell, 79, 1157–1164. Hood, J.D., Bednarski, M., Frausto, R., Guccione, S., Reisfeld, R.A., Xiang, R. and Cheresh, D.A. (2002) Science, 296, 2404–2407. Gurrath, M., Müller, G., Kessler, H., Aumailley, M. and Timpl, R. (1992) Eur. J. Biochem., 210, 911–921. Haubner, R., Gratias, R., Diefenbach, B., Goodman, S.L., Jonczyk, A. and Kessler, H. (1996) J. Am. Chem. Soc., 118, 7461–7472. Haubner, R., Wester, H.-J., Reuning, U., Senekowitsch-Schmidtke, R., Diefenbach, B., Kessler, H., Stöcklin, G. and Schwaiger, M. (1999) J. Nucl. Med., 40, 1061–1071.

54 (a) For the chemical synthesis of Echistatin, see: Garsky, V.M., Lumma, P.K., Freidinger, R.M., Pitzenberger, S.M., Randall, W.C., Veber, D.F., Gould, R.J. and Friedman, P.A. (1989) Proc. Natl. Acad. Sci. U.S.A., 86, 4022–4026; (b) For the structural elucidation by NMR, see: Saudek, V., Atkinson, R.A. and Pelton, J.T. (1991) Biochemistry, 30, 7369–7372. 55 (a) Liu, B.A., Jablonowski, K., Raina, M., Arcé, M., Pawson, T. and Nash, P.D. (2006) Mol. Cell, 22, 851–868; (b) Machida, K. and Mayer, B.J. (2005) Biochim. Biophys. Acta, 1747, 1–25. 56 (a) Overduin, M., Rios, C.B., Mayer, B.J., Baltimore, D. and Cowburn, D. (1992) Cell, 70, 697–704; (b) Waksman, G., Kominos, D., Robertson, S.C., Pant, N., Baltimore, D., Birge, R.B., Cowburn, D., Hanafusa, H., Mayer, B.J., Overduin, M., Resh, M.D., Rios, C.B., Silverman, L. and Kuriyan, J. (1992) Nature, 358, 646–653; (c) Booker, G.W., Breeze, A.L., Downing, A.K., Panayotou, G., Gout, I., Waterﬁeld, M.D. and Campbell, I.D. (1992) Nature, 358, 684–687. 57 Ogura, K., Shiga, T., Yuzawa, S., Yokochi, M., Burke, T.R. and Inagaki, F., unpublished results: PDB access code: 1X0N. 58 Gao, Y., Voigt, J., Wu, J.X., Yang, D. and Burke, T.R. Jr. (2001) Bioorg. Med. Chem. Lett., 11, 1889–1892. 59 Shi, Z.-D., Lee, K., Liu, H., Zhang, M., Roberts, L.R., Worthy, K.M., Fivash, M. J., Fisher, R.J., Yang, D. and Burke, T.R. Jr. (2003) Biochem. Biophys. Res. Commun., 310, 378–383. 60 Shi, Z.-D., Wei, C.-Q., Lee, K., Liu, H., Zhang, M., Araki, T., Roberts, L.R., Worthy, K.M., Fisher, R.J., Neel, B.G., Kelley, J.A., Yang, D. and Burke, T.R. Jr. (2004) J. Med. Chem., 47, 2166–2169. 61 Dekker, F.J., de Mol, N.J., Fischer, M.J. E., Kemmink, J. and Liskamp, R.M.J. (2003) Org. Biomol. Chem., 1, 3297–3303.

j27

j29

2 Macrocycles by Ring-closure Metathesis Joe¨lle Prunet, Anderson Rouge dos Santos, and Jean-Pierre Fe´re´zou

2.1 Introduction

This chapter concerns the preparation of macrocyclic products by ring-closing metathesis (RCM) or related processes combining RCM with other types of metathesis, starting from suitably substituted diene, ene-yne, or diyne precursors. Macrocyclic rings of 10 or more members have been taken in consideration. Such macrocycles can exist individually or as part of a polycyclic system of the bridged, fused, or ansa type. This review is not intended to be comprehensive, but should be regarded as a critical overview of the scope and limitations of RCM as applied for the formation of macrocycles. We have tried to adopt a plan that reﬂects the different types of problems/solutions encountered throughout the huge number of methodological or synthetic studies devoted to the application of macrocycle ring-closing metathesis (M-RCM) to the construction of macrocycles during the past decade. Some case studies that involve particularly biologically and/or structurally interesting families of compounds for which, in some cases, several synthetic and methodological RCM strategies have been developed, will be presented in the last part of the review. One of the main – and also more complex – areas of application of M-RCM is the ﬁeld of natural product chemistry [1]. Natural macrocycles frequently show remarkable pharmacological properties, exhibiting, for example, antibiotic, immunosuppressant, antifungal, and anticancer activities, some of them being potent inhibitors of speciﬁc enzymes associated with several target diseases [2]. The main types of skeletons encountered in natural macrocycles include diterpenes, macrolactones, macrolactams, macroketides (macrocyclic polyketides), or cyclic peptides inter alia. This unique proﬁle continues to inspire huge efforts for their total synthesis and for the development of more speciﬁc or simpliﬁed pharmacologically targeted analogs. This is especially the case for natural macrocycles (such as those of marine origin)

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

30

j 2 Macrocycles by Ring-closure Metathesis which are difﬁcult to access and only available in minute quantities, insufﬁcient for pharmacological trials. Synthesis is also sometimes the only way to determine their chemical structure with certainty. Whatever the object of performing such complex synthetic tasks, one common feature of most of these natural macrocyclic molecules is their complex structure, which often includes various substituents, polar functional groups and/or unsaturations on a skeleton bearing several stereogenic centers. The construction of these molecules requires speciﬁc chemo-, regio-, and stereoselective reactions among which a key one is the formation of the macrocycle itself, generally in one of the late steps of the synthesis, when the required acyclic precursor already comprises most of these complex functionalities. M-RCM has quickly emerged as one of the most useful tools for such a decisive purpose because of its ease of operation and its excellent functional group tolerance. These advantages are quite obvious when considering the spectacular speed at which it is supplanting classical well-established key macrocycle-closing methodologies such as macrolactonization, macrolactamization, or Wittig/Horner-Emmons reactions. Furthermore, application of M-RCM is not restricted to complex natural products or their pharmacologically signiﬁcant analogs. Alone, or in combination with other metathesis or non-metathesis processes, it has also been applied to the construction of several non-natural macrocyclic substrates such as catenanes, rotaxanes, etc. Complex strategies involving M-RCM as a key step in various tandem, domino, or even some multicomponent reaction chemistry (RCM) processes are also appearing, offering a promising entry into the diversity-oriented construction of macrocycles. The advent of RCM [3] as an essential tool for the synthesis of macrocycles undoubtedly arose from two decisive factors: . The development of stable metal-carbene precatalysts (called catalysts for sake of simplicity throughout this review) – easy to handle and exhibiting exceptional functional-group tolerance. Nowadays, the ruthenium catalysts are the most popular. . The second factor, as a practical consequence of the ﬁrst, has been the early commercial availability of some of these catalysts, particularly the ﬁrst generation catalyst G1 in 1996 and, some years later (1999), the second generation catalyst G2 (Figure 2.1). These catalysts are now routinely used in organic chemistry laboratories. Historically, two remarkable well-deﬁned families of complexes were designed in the early 1990s. The ﬁrst-generation metal-carbene of practical synthetic utility was the very active molybdenum catalyst S1 developed by Schrock and coworkers in 1990 [4]. A little while after this, the ﬁrst-generation stable ruthenium ﬁve-coordinate 16-electron alkylidene(II) complex G0 was prepared by Nguyen and Grubbs, showing rather low reactivity [5]. The Grubbs research group succeeded in developing a practical large-scale production of the more reactive bis(PCy3)ruthenium alkylidene complex G1 by substituting the Ru¼CHCH¼CPh2 carbene unit by the Ru¼CHPh moiety [6]. This second ﬁrst-generation carbene rapidly became universally used for a huge number of synthetic applications.

2.1 Introduction

iPr

iPr N Me Me Ph

(CF3)2MeCO Mo

1990 Commercial highly reactive, sensitive to O2 and H2O

(CF3)2MeCO S1

Schrock

PCy3 First generation Grubbs type ruthenium catalysts

Cl Cl G0

Ru

Ph Ph

PCy3

1992 The first ruthenium catalyst Second generation Grubbs type ruthenium catalysts Cy = cyclohexyl Mes = mesityl

N Mes Mes N Ph Cl Ru Cl H PCy3 GN 1999 Grubbs/Nolan catalyst highly reactive

PCy3 Cl Cl G1

Ru

Ph H

PCy3

1996 Commercial effective first-generation N Mes Mes N Ph Cl Ru Cl H PCy3 G2 1999 Commercial Grubbs second-generation highly reactive

Figure 2.1 The most popular catalysts for RCM.

However, the lack of reactivity of G1 with tri- and tetrasubstituted oleﬁns led the Grubbs group to develop a second generation of catalysts bearing N-heterocyclic carbene ligands, with a remarkable metathesis activity together with an improved stability. Among these new super-catalysts was the ruthenium complex GN bearing a bis-mesitylimidazol-2-ylidene ligand, which was almost simultaneously developed by the Grubbs [7] and Nolan [8] groups. Subsequently, the catalyst G2 with a bismesityl-4,5-dihydroimidazol-2-ylidene ligand [9] appeared, and received more attention from synthetic chemists. Nowadays, the commercial second-generation catalyst G2 is routinely utilized by synthetic organic chemists. Most of the recent synthetic applications of M-RCM involve one of the above catalysts, particularly G1 or G2, chosen as a function of its own reactivity proﬁle, generally after preliminary reaction assays on the genuine substrate or speciﬁc model compounds. The sensitivity of the RCM reaction to steric hindrance is well established. These ruthenium catalysts exhibit high afﬁnity for carbon-carbon double bonds and are compatible with the presence of many functional groups, even the presence of free polar hydroxyl or amino groups. Their use does not require special conditions such as glove boxes, which are required when using Schrocks molybdenum catalyst.

j31

32

j 2 Macrocycles by Ring-closure Metathesis 2.2 How to Cyclize?

From the huge number of model studies and synthetic applications realized during the last ten years, it is possible nowadays to delineate the main factors that affect the course of M-RCM. The following section, through convenient examples, illustrates some of the most representative studies that have helped to establish the scope and limitations of the application of RCM to the synthesis of macrocycles. These contributions have made the use of RCM a predictive tool particularly well suited for the closing key step. However, as will be seen later, some complex situations can still occur where the course of the RCM reaction remains unpredictable and continues to require further studies. Two main questions are to be asked when envisioning M-RCM as a key step in a synthetic strategy: . What are the best conditions to perform the cyclization reaction? . Is it possible to orientate the reaction to get an exclusive E or Z double bond accordingly? This second – more complex – point only arises when the ﬁnal synthetic target contains the newly formed double bond itself or one stereoselective transformation adduct of it. However, application of RCM to the synthesis of macrocycles is facilitated by the fact that many target macrocycles do not include double bonds or are constructed using synthetic strategies that design the closing reaction at a sigma bond. In both cases, the initial RCM closing step is followed by a subsequent hydrogenation reaction that obviously closes the stereochemical debate. Many reviews deal with the main mechanistic aspects of the metathesis reaction [10]. There are three basic metathesis reactions (apart from polymerization reactions): the ring-closing metathesis (RCM), the cross-metathesis (CM) and the ringopening metathesis (ROM) [11]. Among the fundamental aspects that govern the reaction course, the thermodynamic versus kinetic issue is particularly important when considering the application of RCM to the construction of macrocycles. Basically, the metathesis reaction consists of reversible [2 þ 2]-cycloaddition processes between two unsaturated partners, leading to equilibrium mixtures where kinetic, thermodynamic, and entropic parameters are decisive. The main steps are summarized in Scheme 2.1. 2.2.1 A Thermodynamic versus Kinetic Issue

As can be seen from Scheme 2.1, under appropriate equilibrating conditions, the macrocycle will be produced as the thermodynamic product. This clearly appears in the following example (Scheme 2.2). During the synthesis of the 14-membered lactam ﬂuvirucin B1 by Hoveyda and coworkers from diene 1, only the dimer 2 resulting from the head-to-head coupling between the less substituted double bond

2.2 How to Cyclize?

H2C

[M] CH2

Macrocycle (thermodynamic)

CH2

Oligomers

Precursor diene

H2C

[M]

[M] [M] CH2

H2C

CH2

CH2

Dimer (generally kinetic)

H2C CH2

[M]

Scheme 2.1 Mechanism of RCM.

of the dienic substrate is formed when using G1 catalysis [12]. With the more reactive molybdenum catalyst S1, the cyclic precursor 3 of ﬂuvirucin is produced in excellent yield. Interestingly, direct treatment of the preceding dimer 2 (the kinetic product of metathesis) with S1 also furnished 3 (the thermodynamic product of metathesis), which was subsequently transformed into the aglycon of ﬂuvirucin B1 (Sch 38516). Conversion of a head-to-head dimer to the corresponding macrocycle has also been effected during the synthesis of cylindrocyclophane precursors by Smith and coworkers (see Section 2.5.1). A similar reaction proﬁle was observed by Hagiwara et al. when synthesizing muscone from citronellal [13]. When diene 4 (R¼Me) is exposed to catalyst G1, only the corresponding dimer is formed (Scheme 2.3). However, with terminal oleﬁn 5

OTBS O

Et

TBSO

G1 only product

Et

OTBS O

Et

NH

HN

O

Et Et

Et

HN

Homodimer 2

1 S1 (20 mol%)

X

C6H6 , 22 °C 90%

S1 (22 mol%) pretreated with ethylene OOH

50-55%

6 5

OTBS O

Et

Et

O

HN

3

O

Et

Et

HN Fluvirucin B1 (Sch 38516)

Scheme 2.2 Kinetics versus thermodynamics: catalyst effect.

OH NH2

j33

34

j 2 Macrocycles by Ring-closure Metathesis O

O G1 R R

4 5

CH2Cl2 45 °C, 21 h

6 0% (dimer: 83%) 78% (E/Z mixture) + 18% dimer

R = Me R=H O H2, 10% Pd/C EtOH, 98%

(R)-(-)-Muscone Scheme 2.3 Kinetics versus thermodynamics: inﬂuence of the oleﬁn substitution.

(R¼H), the desired product 6 is obtained in excellent yield. By lessening the steric hindrance of the oleﬁn, the macrocyclic compound can be reached with G1, showing once again that this catalyst is unable to react with hindered oleﬁns. On the other hand, the alcohol corresponding to 4 (68 in Scheme 2.26), does cyclize with catalyst G2 (see Section 2.5.5). 2.2.2 General Experimental Conditions

We have seen that the appropriate choice of the catalyst and/or the oleﬁn substitution pattern can dramatically shift the equilibrium in favor of the macrocyclic product. Other factors are important as well, such as the general experimental conditions. Danishefsky and coworkers have highlighted the effects of solvent, temperature, and concentration on the yield of macrocycle RCM for the synthesis of analogs of epothilone 490 [14]. As shown in Scheme 2.4, RCM of compound 7 in

10

O HO

N

O OR R = TES

O

N

G2

O 7

O

13

HO O

(10-20 mol%)

O

OR

O

8

Solvent

Temp.

Time

Conc.

Yield

CH2Cl2 PhMe PhMe

40 °C 110 °C 110 °C

6h 25 min 15 min

2 mM 2 mM 0.5 mM

<20% 30% 65%

Scheme 2.4 Temperature and solvent effects.

2.2 How to Cyclize?

reﬂuxing dichloromethane affords only a poor yield of macrocycle 8. When the reaction is performed in toluene at 110 C, the yield slightly increases. At higher temperature and dilution, the yield improves to 65%. 2.2.3 Inﬂuence of Polar Complexing Groups 2.2.3.1 A Decisive Factor for Success A key factor for the success of M-RCM is the presence of a properly positioned polar complexing group in the dienic substrate, which serves as an anchor group for the catalyst, as shown by Fürstner and Langemann [15]. Metathesis of hexadeca-2,15diene with catalyst G0 in dichloromethane does not lead to the desired 14-membered oleﬁn 9, but only to oligomers (Figure 2.2). When an ester function is present, compound 10 is obtained in 52% yield under the same conditions. However, the position of the carbonyl moiety relative to the oleﬁn is of crucial importance: lactone 11 cannot be obtained by RCM with G0, while its regioisomer 12 is produced in good yield (Figure 2.2). This difference in behavior is explained in Figure 2.3. If the ruthenium is too tightly complexed by the polar group (such as in 5- or 6-membered chelates), the metathesis reaction can be inhibited. A thorough study by Weiler et al. [16] also illustrates this point: in the case of n ¼ 1 or 9, 5- and 6-membered chelates, respectively, are formed as intermediates, and the RCM yields of lactones or lactams 14 are poor (Scheme 2.5). For n ¼ 5, the chelate is a 10- or 11-membered ring, which is a favorable case for RCM (see Figure 2.3).

O

O O

O 9 0%

N Boc

O 11 0%

10 52%

Figure 2.2 Inﬂuence of polar groups.

O O

[M]

[M]

O

O OR

OR

M L L Favorable case

O

Unfavorable cases

Figure 2.3 Possible chelation.

N Boc 12 84%

j35

36

j 2 Macrocycles by Ring-closure Metathesis O

O ( )n

G1 (2 mol%) CH2Cl2, 2 mM

X ( ) 10-n

( )

13

X ( )

n

10-n

14 Yield Product Start. mat.

X

n

Hours

O

1 5 9

30 " "

11% 62% 0%

41% 11% 66%

NH

1 5 9

24 1 26.5

0% 87% 7%

82% 12% 83%

NBoc

1 5 9

24 8 24

0% 62% 20%

93% 11% 67%

Scheme 2.5 Importance of polar group localization.

2.2.3.2 The Titanium Trick Fürstner and Langemann devised a means of disrupting the chelate that is compatible with the metathesis catalysts. By using a mild Lewis acid, Ti(OiPr)4, which is complexed by the carbonyl group and thus releases the ruthenium, RCM of esters 15 proceeds in 79–80% yield (Scheme 2.6) [17]. No trace of the desired macrocycles 16 was observed in the absence of Lewis acid. Compound 16 (R¼SiMe2tBu) is then converted to gloeosporone by KMnO4 oxidation of the oleﬁn into the corresponding diketone and deprotection of the silyl ether.

H G1 (3 mol%)

H OR O

O ( )

H

4

15 R = Bn R = SiMe2tBu

H

OH

H O

O

O ( )

H Scheme 2.6 The titanium trick.

4

O ( )

79% (E/Z 3.6:1) 16 80% (E/Z 2.7:1) No reaction without Lewis acid O

R = SiMe2tBu

OR O

Ti(OiPr)4 (30 mol%) CH2Cl2, 40 °C

Gloeosporone

4

2.2 How to Cyclize?

Another example deals with the synthesis of 6-nor-analogs of ﬂuvirucin reported by Bracher and Baltrusch [18], where the yield of the RCM step improves from 0% to 72% by addition of 30 mol% of titanium isopropoxide. 2.2.3.3 A Particularly Favorable Case: The Template-Directed RCM One speciﬁc case of favorable conformational beneﬁt for a RCM is the templatedirected RCM where non-covalent interactions around an appropriate acidic and/or metal template serve to favor productive conformations. This has been found to be useful for linear substrates devoid of intrinsic conformational constraints [19]. RCM of ether 17 in the presence of lithium ions gives the best yield of product 18, in accordance with the preferential binding of Liþ by [12]crown-4, which resembles diene 17 (Scheme 2.7). 2.2.4 Chemoselectivity

Oleﬁn metathesis is very sensitive to steric hindrance, and in particular to the substitution pattern of the reacting oleﬁns. In the course of their total synthesis of the ﬁrst proposed structure of amphidinolide A, Maleczka et al. nicely illustrated this feature [20]. Although the yield is modest and the catalyst loading high, RCM of polyene 19 with catalyst G2 furnished only macrocycle 20 (Scheme 2.8), by preferential reaction of the terminal oleﬁns versus the 1,1- and 1,2-disubstituted and trisubstituted alkenes. It is worth noting that the less reactive ﬁrst-generation Grubbs catalyst G1 only truncated the allylic alcohol to furnish the corresponding methyl ketone [21]. In the case of an approach toward cytotrienins, Panek and coworkers achieved different chemoselectivity according to the catalyst they employed [22]. They ﬁrst tried RCM of polyene 21 with catalyst G2 (Scheme 2.9). Unexpectedly, only the macrocyclic conjugated diene 22 is obtained; presumably, initiation of the metathesis reaction occurs at the internal bond of one of the dienes. On the other hand, use of catalyst G1 leads to the expected conjugated triene 23 in 73% yield.

O O

G1 (5 mol%) O

O 17

=M

5 equiv salt CH2Cl2 - THF 0.02 M, 45 oC, 1 h No salt LiClO4 NaClO4 KClO4

Scheme 2.7 Template-directed metathesis.

O O

O O 18

39% (Z/E 38:62) >95% (Z/E 100:0) 42% (Z/E 62:38) 36% (Z/E 36:64)

j37

38

j 2 Macrocycles by Ring-closure Metathesis OH

OH PMBO

G2 50 mol%

PMBO O

TBSO 19

CH2Cl2 40 °C, 24 h 35%

O

O

PMBO PMBO O

TBSO

O

O

20 OH HO

HO

Presumed amphidinolide A O

HO

O

O

Scheme 2.8 Highly chemoselective RCM in the synthesis of presumed amphidinolide A.

OMe

NH

G2 (10 mol%) CH2Cl2, 40 oC

TIPSO

OMe 47%

NH OMe O

TIPSO

21

O OMe

TBSO 22

G1 (20 mol%) CH2Cl2, 40 oC

OMe

OMe

TBSO

MeO

NH

73%

TIPSO TBSO

OMe O OMe

23 Macrocyclic core of cytotrienins

Scheme 2.9 Effect of catalyst on chemoselectivity.

2.2.5 Substrate ‘‘Tuning’’

Among the factors affecting the course of the RCM reaction, those depending on substituent conﬁguration and/or functional group protection are among the most difﬁcult to anticipate. Several cases illustrate the important inﬂuence on the course of the reaction of apparently insigniﬁcant structural modiﬁcations to the precursor backbone. 2.2.5.1 Conﬁgurational/Conformational Aspect One single inverted stereocenter can dramatically affect the course of the metathesis reaction. Schreiber and his group obtained substrates 24 and 25 by acylation of a

2.3 Factors Inﬂuencing the Double-Bond Conﬁguration

O HH

O

H

O

HN

MeO

O

G2 Ph (15 mol%)

N O

HH H

63% MeO

N

O

Ph

24

O

HN

Ph

N H MeO

Ph

G2 (15 mol%)

O N Ph

O

26

O O

HN

N

O

HH

Ph

O

O

O

Ph

O

N

O N

76%

O O

HN HH

O

O

O

H MeO

25

O

O Ph

N

H

27

Scheme 2.10 Effect of a single stereocenter on the RCM outcome.

chiral amine by a racemic mixture of products of a Ugi 4CC-IMDA sequence, so they differ by the stereocenter marked in Scheme 2.10 (to be rigorously exact, 24 and the enantiomer of 25 are epimeric at this center [23]). In the case of 24, a simple RCM takes place, leading to the 17-membered macrolactone 26. For 25, the reaction follows a totally different pathway: RCM/ROM, forming the seven-membered lactam and simultaneously opening the oxo-bridged bicycle, is accompanied by RCM which leads to the 12-membered lactam 27. The pathway selectivity is very high: none of the product resulting from the alternative pathway is observed in either case. 2.2.5.2 Inﬂuence of Functional Group Protection Cyclic diol protecting groups are not always conducive to a successful RCM. For example, when Banwell and McRae submitted acetonide-protected 1,3-diol 28 to the Grubbs ﬁrst-generation catalyst, none of the desired macrocyle was produced, but cyclohexene 29 was obtained in 81% yield, probably because the acetonide protecting group facilitated interaction of the double bonds of the carboxylic acid portion of the molecule (Scheme 2.11) [24]. To circumvent this problem, substrate 30 was synthesized, where the diol was protected as two silyl ethers, and RCM of this compound led to the desired 18-membered lactone 31 in 70% yield under the same reaction conditions.

2.3 Factors Inﬂuencing the Double-Bond Conﬁguration

A key challenge in M-RCM, especially for the construction of unsaturated targets in which the oleﬁn moiety is installed during the metathesis reaction, is the control of the E/Z selectivity. This applies not only to macrocycles but also to medium-size

j39

40

j 2 Macrocycles by Ring-closure Metathesis OTBS

OTBS

G1 (10 mol%) CH2Cl2 20 °C, 14 h

O O

O O

81%

O

29

28 O OH OTBS OTBS

OH

G1 (10 mol%) CH2Cl2 20 °C, 14 h

OTBS OTBS

70%

O O

O Aspicillin

30

O 31 (E/Z 3:1)

Scheme 2.11 Inﬂuence of protecting groups.

cycles starting from 8-membered rings. All prerequisite requirements described above for M-RCM cyclization to take place were based on more-or-less objective rationales in relation to the structure/conﬁguration of the putative dienic precursor. However, the design of an appropriate substrate leading to the desired conﬁguration of the incipient double bond is far more complex and therefore difﬁcult to anticipate. In addition, several factors can also inﬂuence the stereoselectivity of the RCM reaction [25]. 2.3.1 A Thermodynamic versus Kinetic Issue

As we have seen before, the RCM reaction is generally under thermodynamic control. The inﬂuence of the nature of the catalyst on the RCM selectivity was observed by Grubbs et al. (Scheme 2.12) [26]. By submitting 32 to a more active

G1 or G2 CH2Cl2 (3 mM)

O ()

O

8

() 2

R

O

45 oC

32

O 33

Entry

R

1 2 3 4 5 6

H " Et (Z) " (CH2)4Me (Z) CH2OAc (E)

Cat. (mol%) Time

Yield

E/Z

G2 G1 G2 G1 G2 "

100% 97% 100% 77% 100% 80%

11.5:1 4.5:1 9.7:1 4.5:1 10.8:1 9.7:1

(1.0) (5.0) (0.5) (5.0) (0.5) (2.0)

40 min 5h 30 min 6.5 h 40 min 3h

Scheme 2.12 Inﬂuence of the nature of the catalyst on the E/Z ratio.

2.3 Factors Inﬂuencing the Double-Bond Conﬁguration

GN CH2Cl2, 40 °C O O 34

R=H O

R

nPr O

PCy3

O O Z-35

Cl Cl

O nPr

F1

86%, 1 isomer aq. HCl

O

O

E-35

nPr O

Ph

PCy3

O

O F1 CH2Cl2, 40 °C

Ru

R

R = H 78% (E/Z 7.7:1) R = OMOM 79% (E only)

HO HO

O

R

nPr O Herbarumin I R = H 90% Herbarumin II R = OH 84%

Scheme 2.13 E/Z selectivity by choice of the catalyst: synthesis of herbarumin I and II.

catalyst, a higher proportion of the E-isomer of lactone 33 is produced. The E/Z ratio does not signiﬁcantly evolve with time when using catalyst G1, while it increases from 4.8:1 at 20% conversion to 11.5:1 at 100% conversion with G2. The authors also reported that a 4:1 E/Z mixture of lactone 33 can be transformed into a 12:1 mixture of E/Z isomers with G2, while the ratio does not change in the presence of G1. This increase in the E/Z ratio may be due to secondary metathical isomerization with the more active catalyst G2, ultimately leading to the thermodynamic equilibrium of the ring-closure product. In some cases, the proper selection of the catalyst can lead selectively either to the thermodynamic or the kinetic metathesis product [27]. Fürstner et al. nicely illustrated this trend during their synthesis of herbarumin I and II from ester 34 [28] (Scheme 2.13). Semiempirical calculations showed that model lactone Z-35 is ca. 3.5 kcal mol1 more stable than the corresponding lactone E-35 (R¼H), suggesting that thermodynamic control would likely give the undesired isomer for the synthesis of the target molecules. This prediction was veriﬁed when compound 34 was submitted to the second-generation complex GN: the more stable Z-35 lactone is produced as the only isomer. When catalyst F1 is employed, whose activity is comparable to that of ﬁrst-generation Grubbs catalyst G1, the E-isomer is obtained as the major product. The E/Z ratio does not evolve with time, proving that the reaction is under kinetic control with this catalyst. Herbarumin II precursor E-35 (R¼OMOM) is also synthesized with an excellent selectivity using F1. A similar observation was made by Marco and coworkers when synthesizing microcarpalide [29]. In the case of the cyclization of acetonide 36, the Z-oleﬁn is the thermodynamic product, produced with the second-generation catalyst G2, while use of G1 gives a 2:1 mixture of E/Z alkenes 37 (Scheme 2.14). These compounds could be separated by chromatography and the E-isomer was subsequently converted to microcarpalide by removal of the MOM and the acetonide protecting groups.

j41

42

j 2 Macrocycles by Ring-closure Metathesis OMOM O

G2 (20 mol%), 1 mM CH2Cl2, 40 °C, 1 h

O

O

O Z-37 73% (Z only)

OMOM O

OMOM O

O 36

O

O G1 (20 mol%), 1 mM CH2Cl2, 40 °C, 24 h

O

O

Microcarpalide

O 37 67% (E/Z 2:1)

Scheme 2.14 Orientation of the E/Z-selectivity by choice of the catalyst: synthesis of microcarpalide.

2.3.2 General Experimental Conditions

The experimental conditions can also affect the stereoselectivity of the M-RCM reaction. For example, the key step of the synthesis of recifeiolide from ester 38 with catalyst G0 by Fürstner and Langemann proceeds in good yield, but the E/Z ratio depends on the temperature of the reaction (Scheme 2.15) [15]. Nakagawa et al. also reported a variation of the E/Z ratio with the reaction temperature during the formation of azacycloundecenes in the presence of G1: it changes from 12.5:1 at 20 C to 5.2:1 at 50 C [30]. In the reaction shown in Scheme 2.15, the selectivity could also depend on the solvent, as it is the case for 39, used for the preparation of the 17-membered diether 40 using catalyst G1 by Tori et al.: in ether, the macrocyclic alkene is obtained as a 1:1 mixture of E- and Z-oleﬁns, and in dichloromethane, the E-isomer of 40 is produced exclusively (Scheme 2.16) [31]. No rational explanation is given for the effect of solvent or temperature. 2.3.3 Substrate ‘‘Tuning’’

In the course of total synthesis of highly functionalized targets, protecting groups are easily modiﬁable. They often have a dramatic inﬂuence on the stereochemical O O

O G0

38

O

Recifeiolide CH2Cl2, 40 °C Toluene, 80 °C

80% (E/Z 4.7:1) 77% (E/Z 3.3:1)

Scheme 2.15 Inﬂuence of temperature.

2.3 Factors Inﬂuencing the Double-Bond Conﬁguration

O

O G1 20 °C

O 40

O

Et2O CH2Cl2

39

52% (E/Z 1:1) 69% (only E)

Scheme 2.16 Inﬂuence of solvent.

outcome of RCM, especially if a cyclic protecting group is employed. When designing the retrosynthesis of tonantzitlolone, Kirschning and his group envisaged an intermediate with an oleﬁn moiety at C4–C5, which could be Z or E, as it was to be subsequently oxidized to a hydroxy ketone [32]. Diol 41 was ﬁrst protected as the corresponding acetonide 42, which furnished the Z-oleﬁn 43 by RCM with good stereoselectivity (Scheme 2.17). Unfortunately, dihydroxylation of this alkene proved

OTBS

HO HO

O H 41

OH

O O 42

5

H 44

G2 (10 mol%) CH2Cl2 (2.0 mM) 40 °C

4

OH

O

HO

O

O

O

HO

H

O

O

HO

O

HO

O

H 43 70% (Z/E ~ 6:1)

H 45 72% (E/Z ~ 5:1)

X

Oxidation at C4-C5 impossible

(ent)-Tonantzitlolone

Scheme 2.17 Inﬂuence of a diol cyclic protecting group on the stereochemical outcome of RCM.

j43

44

j 2 Macrocycles by Ring-closure Metathesis 12

12

S 13

RO

N

7 5

3

Y

X

G1 (50 mol%) 1 mM

13

RO

N

7

Benzene 20 °C, 24 h

O O

46a X = α-OTBS, Y = O, R = TBS 46b X = β-OH, Y = α-OTPS, R = TBS

S

5

3

Y

X

O O

47a 86% (E/Z 1:1.7) 47b 81% (E/Z 9:1) O S

47a

HO

N O O

OH O

Epothilone A

Scheme 2.18 Inﬂuence of remote substituents.

impossible. However, cyclization of keto diol 44, derived from the same precursor 41, led to the E-isomer 45, which could be transformed into the enantiomer of tonantzitlolone. Presumably, the conformational constraint induced by the acetonide prevents the formation of the E-oleﬁn, which is the more stable product in the absence of diol protecting group. Precursors of epothilone A (with a Z-12,13-double bond which will lead to the required cis 12,13-epoxide) ﬁgure among the most studied macrocycles formed by RCM, and they are good examples of the inﬂuence of remote substituents on the E/Zselectivity. In the synthesis by Nicolaous group, the best proportion of Z-isomer is 1:1 [33]. In a similar approach toward this natural product, Danishefsky and coworkers showed that substituents far away from the reaction site (at C3, C5, or the C7 hydroxyl protecting group) have a profound inﬂuence on the selectivity of the alkene formation (Scheme 2.18) [34]. The best result in view of the epothilone A synthesis was a 1:1.7 E/Z ratio in the case of 46a, which was eventually converted into the target molecule. For 46b, which is epimeric at C3 and possesses a protected alcohol instead of a ketone at C5, the E/Z ratio switches to 9:1. It seems that there is no reliable method to predict the outcome of this RCM reaction. Alternative strategies for this total synthesis include either formation of the Z-alkene on a linear precursor and macrolactonization to close the 16-membered ring [35] or ring-closing alkyne metathesis followed by partial reduction of the newly created triple bond (see Section 2.3.4). Other examples of the inﬂuence of remote substituents on the stereoselectivity of RCM will be detailed in Section 2.6.1. 2.3.4 Solutions

As we have just seen, oleﬁn metathesis is a powerful method to synthesize large rings containing alkenes, but the E/Z selectivity is variable and rarely predictable. The answer may lie in an indirect methodology reported by Fürstner and coworkers.

2.4 Ene-yne M-RCM

S RO

N O O

OR O 48

[MoN(Ar)3] Toluene CH2Cl2

RO

R

80 °C, 8 h 80%

O 49

R = TBS

O

OR O

H2, CH2Cl2 Lindlar, quinoline 100% N

Mo

N

S

N

Z

RO Epothilone A [MoN(Ar)3]

N

1) HF/CH3CN

O

2) DMDO 50

O

OR O

Scheme 2.19 RCAM: the solution to reliable stereoselectivity for alkene formation in macrocycles.

Ring-closing alkyne metathesis (RCAM) leads to alkyne macrocycles in excellent yields [36]. When combined with hydrogenation using a Lindlar-type catalyst, it is a convenient pathway to macrocyclic Z-cycloalkenes. Fürstner and coworkers utilized this strategy for numerous total syntheses of natural products [37], epothilone A being one of them [38]. The metathesis step is effected with the molybdenum precatalyst [MoN(Ar)3] (Scheme 2.19). In the presence of dichloromethane, the active catalyst (the exact structure of which is unknown) reacts with 48 to give alkyne 49 in 80% yield. Hydrogenation of 49 exclusively leads to Z-oleﬁn 50. It is interesting to note that the oleﬁn metathesis reported by Nicolaou with the diene analog of 48 gives a 1:1 ratio of E/Z isomers [33]. The macrocyclic alkynes can also be reduced in a two-step fashion (E-selective Rucatalyzed hydrosilylation followed by desilylation with AgF) to furnish E-cycloalkenes with selectivities ranging from 90:10 to >98:2, as described by Fürstner and Radkowski [39].

2.4 Ene-yne M-RCM

Along with diene and diyne metathesis, ene-yne metathesis has also been employed to form macrocycles. This type of metathesis is performed with the catalysts used for oleﬁn metathesis, and the yields are improved in the presence of ethylene, which forms the highly reactive [Ru]¼CH2 species. Shair and coworkers took advantage of this reaction twice in the course of their total synthesis of longithorone A [40]. When ene-ynes 51 and 52 are treated with ruthenium complex G1 under an atmosphere of

j45

46

j 2 Macrocycles by Ring-closure Metathesis

MeO G1 (50 mol%)

MeO

OTBS OTBS

H2C CH2 CH2Cl2, 40 °C

OTBS OTBS OTBS

31% 53 TBSO E/Z > 25:1

51

Longithorone A

G1 (50 mol%) TBSO

OMe

TBSO

H2C CH2

OMe

CH2Cl2, 40 °C OTBS

42%

OTBS 54 E/Z > 25:1

52

Scheme 2.20 Ene-yne RCM in the synthesis of longithorone A.

Intramolecular pathway: [Ru] CH2 ()

0-4

[Ru] ()

()

0-4

0-4

1,2-disubstituted Intermolecular pathway:

R1

[Ru] CH2

R1

[Ru]

H2C CH2

R2

R1 R2 1,3-disubstituted

Scheme 2.21 Intramolecular and intermolecular ene-yne metathesis pathways.

ethylene, macrocycles 53 and 54 are produced as single atropdiastereomers, and with a E/Z selectivity better than 25:1 (Scheme 2.20). Both compounds bear a 1,3-disubstituted diene, proving that RCM occurs through the intermolecular ene-yne metathesis pathway (Scheme 2.21), which is usually the case for rings of size 12 or greater.

2.5 Tandem Processes Involving M-RCM

A few examples of tandem processes have been designed for the synthesis or the transformation of natural products.

2.5 Tandem Processes Involving M-RCM

2.5.1 Tandem CM/RCM

The Smith group reported a very elegant synthesis of cylindrocyclophanes A and F [41]. Dimerization of bis-oleﬁn 55 with ﬁrst-generation Grubbs catalyst G1, followed by ring-closing metathesis of the resulting dimer furnishes exclusively the [7,7]paracyclophane 56 in 61% yield as a single (E,E)-isomer (Scheme 2.22). The yield is not improved with catalyst G2, but use of Shrock molybdenum catalyst S1 gives optimum result (72% yield). None of the head-to-head dimerization products are observed during this reaction. In a very interesting experiment, these dimers, prepared by another route, were independently submitted to catalyst S1, and compound 56 was obtained as the exclusive product in each case. The [8,6]-cyclophanes were not detected. The dimers probably undergo a cascade of metathesis reactions leading to the most stable product [7,7]-paracyclophane 56. This intermediate was subsequently converted to cylindrocyclophane F by hydrogenation of the oleﬁns and deprotection of the methyl ethers. 2.5.2 Tandem ROM/RCM

Lazarova et al. described the ring contraction of 16-membered ring macrolide antibiotics such as josamycin (57) to their 14-membered analogs [42]. In the presence of

MeO MeO

OMe

MeO

OMe

OMe

55

56

G1, CH2Cl2, 20 °C, 75 h G2, benzene, 40 °C, 27 h S1, benzene, 20 °C, 2 h

61% 58% 72%

(E,E only)

R

2) BBr3 84% (2 steps)

OH

HO

1) H2, Pd/C HO

OH R

(-)-Cyclindrocyclophane F R = H (-)-Cyclindrocyclophane A R = OH Scheme 2.22 Tandem CM/RCM process in the synthesis of cylindrocyclophanes.

j47

48

j 2 Macrocycles by Ring-closure Metathesis X Y

X Y CHO HO

MeO O O

O

CHO NMe2 O O

OH O

OR1

57 X = H, Y = OH, R1 = COMe, R2 = Me R2 58 X = H, Y = OH, R1 = H, R2 = H 59 X = OH, Y = H, R1 = COMe, R2 = Me

GN (20 mol%)

1-Hexene (0.2 equiv) O CH2Cl2, 40°C, 24 h O

MeO O O

OSugar OR1

60 78% 61 65% 62 42%

Scheme 2.23 Tandem ROM/RCM process in the synthesis of 14-membered macrolides.

catalyst G2 and 1-hexene, the 16-membered macrocycles 57–59 are cleanly converted to their derivatives 60–62 by a ring-opening/ring-closing metathesis sequence (Scheme 2.23). Catalyst G1 is ineffective for this transformation. The 1-hexene additive is a non-gaseous ethylene equivalent: it allows the formation of the [Ru]¼CH2 complex both from the original G2 catalyst and from the conjugated carbene released after the ring-contraction operation. Titanium isopropoxide, which probably disrupts the chelate between the carbene and the lactone carbonyl moiety (see Section 2.2.3), can serve as an alternative to 1-hexene. Without these additives, the reaction is stoichiometric in ruthenium complex, and the percentage conversion is proportional to the amount of catalyst used. 2.5.3 Tandem RCM/ROM/RCM

During their studies toward the total synthesis of coleophomones, Nicolaou and his group uncovered a stereospeciﬁc RCM transformation of two regioisomeric enol ethers (this reaction will be detailed in Section 2.6.3). Vassilikogiannakis et al. prepared model systems of these enol ethers to try to understand the rationale for this selectivity [43]. During this work, they discovered a dramatic effect of the substitution of one of the oleﬁnic partners on the metathesis outcome (Scheme 2.24). While compound 63a (R¼Me), submitted to catalyst G2, undergoes straightforward RCM reaction to give 11membered ring 64a, the terminal oleﬁn 63b (R¼H) goes through an RCM/ROM/ RCM cascade: after formation of the macrocycle 64b, this ring is reopened and the resulting carbene reacts with the remaining isoprenyl unit to form the spiro compound 65. In the case of 64a, the trisubstitued alkene is too hindered to permit further ring-opening. This ring-opening is possible for 64b, which is ultimately transformed into the thermodynamic product, 5-membered ring 65. 2.5.4 Ring-Expansion Metathesis

This reaction was ﬁrst reported by Grubbs and his group [44]. An ROM/CM sequence between cyclooctene and one of the unsaturated ketones of 66 is followed by

2.5 Tandem Processes Involving M-RCM

R = Me G2 10 mol%

OO

CH2Cl2 40 oC, 3 h

R

OMe O

63%

OO R=H G2 10 mol%

OMe O 63a R = Me (E/Z 1:1) 63b R = H (E/Z 1:1)

CH2Cl2 40 oC, 20 min

64a (E/Z 3:1)

OO

OO 81% 65

64b OMe O

OMe O

[Ru]

Scheme 2.24 Dramatic effect of oleﬁn substitution on the RCM outcome.

O

O

G1 (5 mol%) 5 mM +

66

O

CH2Cl2, 40 °C 53%

67

O

Scheme 2.25 Ring-expansion metathesis reaction.

an RCM to form macrocycle 67. The authors have used other cycloalkenes for this ring-expansion reaction, ranging from cyclopentene to cyclododecene (see Scheme 2.25). 2.5.5 Other One-Pot Multistep Processes

Grubbs and coworkers have also shown that catalysts G1 and G2, treated with hydrogen, afford the corresponding hydride complexes capable of performing hydrogenation reactions [45]. One-pot tandem RCM/hydrogenation or CM/hydrogenation products are easily obtained by pressurizing the reaction ﬂask with hydrogen after the metathesis reaction is complete. Similarly, transfer hydrogenation with isopropanol (for reduction of ketones) or 3-pentanone (for oxidation of alcohols) can be effected by addition of H2 and NaOH to the metathesis-reaction vessel. A 3-step one-pot RCM/transfer hydrogenation/hydrogenation was reported by the same

j49

50

j 2 Macrocycles by Ring-closure Metathesis OH

OH G2 ClCH2CH2Cl 50 °C, 2 d 68

74% conversion (GC) O

O

3-Pentanone, NaOH

H2, pressure

Dehydrogenation by H2 tranfer 76% conversion (GC)

Hydrogenation 100% conversion (GC) 56% overall yield Muscone

Scheme 2.26 Tandem metathesis/hydrogen transfer/hydrogenation sequence.

authors, resulting in a short and efﬁcient synthesis of muscone from alcohol 68 (Scheme 2.26) [45]. 2.5.6 M-RCM as Part of MCR Strategies

In term of diversity-oriented strategies, multicomponent reactions (MCR) represent an attractive and rapid access to libraries of macrocycles inspired by biologically active natural products. Combined with Passerini and Ugi reactions, M-RCM has already shown promising synthetic potential, as illustrated by the pioneering work of Dömling and coworkers [46]. Condensation of isocyanide 69 with carboxylic acid 70 in the presence of paraformaldehyde leads to bis-oleﬁn 71, which is subsequently submitted to RCM in the presence of G1 and titanium isopropoxide to give the 22-membered macrocycle 72 (Scheme 2.27).

2.6 Representative Synthetic Applications

Numerous total syntheses of macrocycles involve RCM as the key step. For some popular targets, the metathesis disconnection has been designed at the same bond by almost all the groups working on their synthesis, so the effects of multiple factors on the outcome of the metathesis reaction can be outlined. A recent review on the metathesis approach to biologically active macrocycles by Van de Weghe and Eustache [47] reports the case of epothilones and peptidomimetic agents, so these compounds will not be mentioned in the following examples.

2.6 Representative Synthetic Applications

N

N

NC O

69 +

O

O

H +

O

O

HO

H

O O

P-3CR (Passerini)

O

NH O

66%

O

O

71

70 N M-RCM, G1 Ti(OiPr)4, CH2Cl2 26%

O O

O

NH O O

O

72

Scheme 2.27 M-RCM in MCR process.

2.6.1 Salicylihalamides/Oximidines: Potent Antitumor Agents with Selective anti-V-ATPase Activity

Salicylic enamides represent a new family of biologically active secondary metabolites of marine origin. These metabolites with promising antitumoral activities are 12-membered lactones, among which salicylihalamide A and oximidine III are two representative leads which have particularly focused the interest of organic chemists (Figure 2.4). An excellent review was written by Yet in 2003 on the chemistry and biology of salicylihalamides, oximidines, and related compounds [48]. Since that year, a few more studies on the synthesis of these compounds have appeared in the literature. In this section, we will try to examine the factors inﬂuencing the E/Z selectivity in

17

H N

H N

17

OH

O

3 1

O O 15

13

O

OH O

OH

O

13

O

9 12 10

Salicylihalamide A Figure 2.4 Salicylihalamide A and Oximidine III.

12 8

N OMe

10

Oximidine III

j51

52

j 2 Macrocycles by Ring-closure Metathesis OPMB

OPMB

OMe O

OMe O OMOM

O 9

73

Catalyst CH2Cl2 20 °C >93%

10

Catalyst G1

G2

OMOM

O

74 Time (h)

E/Z

1.3 6.3 18.3 1.3 6.3 18.3

90:10 88:12 80:20 66:34 68:32 67:33

Scheme 2.28 Inﬂuence of the catalyst: kinetics versus thermodynamics.

the formation of the 12-membered ring of the salicylihalamide derivatives and illustrate the application of relay RCM to the synthesis of oximidine III. 2.6.1.1 Salicylihalamides: Inﬂuence of the Catalyst In the course of their synthesis of salicylihalamide A, De Brabander et al. thoroughly studied the inﬂuence of the catalyst on the outcome of the RCM for 73, which forms the C9–C10 bond [49]. They showed that the ﬁrst-generation catalyst G1 gives better E selectivity than the more active G2 complex (Scheme 2.28). The ratio of E/Z isomers of 74 does not evolve in time with G2, proving that the thermodynamic mixture of products is obtained in this case. On the other hand, kinetic product ratios are probably observed with G1, as the E selectivity decreases with longer reaction times. Maier and coworkers reported the same trend with the similar substrate 75 (Scheme 2.29) [50].

TIPS

TIPS OMe O

OMe O O

75

OMOM

Catalyst CH2Cl2 23 °C

OMOM

O

76

Catalyst

Yield

E/Z

G2 G1

76% 94%

69:31 97:03

Scheme 2.29 Inﬂuence of the catalyst.

2.6 Representative Synthetic Applications

CN

CN

OR O

OR O

3

OPNB

O

G1 CH2Cl2 40 °C, 3 h

77

R=H R = TBDPS R = TBS R = Me

OPNB

O

78

Yield

E/Z

41% 75% 94% 89%

25:75 75:25 87:13 95:05

Scheme 2.30 Inﬂuence of the phenol protecting group with catalyst G1.

2.6.1.2 Salicylihalamides: Inﬂuence of the Phenol Protecting Group The phenol-protecting group has a dramatic effect on the selectivity of the RCM employed to form the 12-membered cycle of salicylihalamides. Georg and his group submitted compounds 77 to catalyst G1: the best E/Z ratio for 78 is obtained with R¼Me, while bulkier protecting groups such as TBS or TBDPS give inferior results (Scheme 2.30) [51]. The unprotected phenol is the only substrate which predominantly leads to the Z isomer. This result was also observed by the same group on cyclic precursors of salicylihalamide A [52]. Fürstner et al. reported the same protecting-group effect using complex GN, with the free phenol 79 exclusively furnishing Z-80 (Scheme 2.31) [53]. They ruled out the interaction of the free phenol with the catalyst, and proved by semi-empirical calculations that there is no correlation between the E/Z ratios and the thermodynamic stabilities of the products, so this effect remains unexplained. The ﬁrst-generation catalyst G1 is unable to react with the trisubstituted oleﬁn at C10.

OPMB

OPMB

OR O

OR O

3

OMOM

O

GN

OMOM

O

Toluene, 80 °C 79

G1 inefficient 10

Time R=H R = TBS R = Me R = MOM

20 h 1h 1.5 h 3h

80 Yield 69% 91% 91% 93%

E/Z 0:100 40:60 66:34 68:32

Scheme 2.31 Inﬂuence of the phenol protecting group with catalyst GN.

j53

54

j 2 Macrocycles by Ring-closure Metathesis CN

CN

OMe O

OMe O OR

O

G1

13

CH2Cl2 40 °C, 3 h

81

OR

O

R=H R = Ac R = MOM R = Piv R = PNB R = TBS

82

Yield

E/Z

44% 87% 84% 94% 89% 94%

97:03 96:04 96:04 96:04 95:05 94:06

Scheme 2.32 Inﬂuence of the alcohol protecting group at C13.

2.6.1.3 Salicylihalamides: Inﬂuence of the Alcohol-Protecting Group at C13 The alcohol-protecting group at C13 has a slight effect on the RCM selectivity for 81, as shown by Georg and coworkers [51], but the E/Z ratios of anisoles 82 are all good (Scheme 2.32). 2.6.1.4 Salicylihalamides: Inﬂuence of the Nature of the C15 Side Chain Several groups (Labrecque [54], Smith and Zheng [55], Yadav and Srihari [56], Georg [51], and Snider [57]) have described total or formal syntheses of salicylihalamide A. Various intermediates 83 have been submitted to complex G1, and the selectivity does not vary much with the nature of the C15 side chain (Scheme 2.33). Once again, the E/Z ratios for 84 are lower when the phenol is protected as a TBS- rather than a Me-ether.

OR1 O R2

OR O R2

O 15

OR3

G1

OR3

O

CH2Cl2 83 R1 Me

TBS

84 R2 (CH2)3OBn (E)-CH2CH=CHCH2OPMB (CH2)2OTBS (CH2)2CN (E)-CH2CH=CHCOOMe (CH2)2CN

R3 TBDPS TBS MOM TBS PNB

Yield

E/Z

Ref

75% 85% 85% 84% 71% 94%

90:10 91:09 90:10 96:04 81:19 88:12

54 55 56 51 57 51

Scheme 2.33 Inﬂuence of the nature of the C15 side-chain.

2.6 Representative Synthetic Applications

In the same fashion, the C15 side chain does not seem to have a great impact on the selectivity observed with G2. Compounds 74 (Scheme 2.28) and 76 (Scheme 2.29) are obtained as 66:34 and 69:31 mixtures of E/Z isomers, respectively, when treating dienes 73 and 75 with G2 in CH2Cl2 at ambient temperature. However, Maier and coworkers reported an example which seemingly shows some deviation from this behavior: if a substrate very similar to 73 (with an extra methylene in the side chain) is reacted with G2 in toluene at 70 C, the selectivity increases to 86:14 [58]. However, in this example the effects of solvent and temperature cannot be ruled out. 2.6.1.5 Salicylihalamides: Miscellaneous We have seen in Section 2.2.5 that the inversion of a single stereocenter of a substrate can orientate the RCM in a totally different fashion. De Brabander et al. [49] veriﬁed this fact when they submitted substrate 85 (whose enantiomer is epimeric at C15 with the natural salicilyhalamide precursors) to RCM: macrocycle 86 is obtained as a 20:80 mixture of E/Z isomers (Scheme 2.34). In comparison, natural salicilyhalamide intermediates 83 (entries 2 and 3 of Scheme 2.33 for example) lead to 90:10 E/Z ratios for 84. Maier and coworkers reported the RCM with G2 of an analog of substrate 75 devoid of substituents at C12 and C13, which leads to a 81:19 E/Z ratio of products [50] (compare with the 69:31 ratio obtained with 75, Scheme 2.29). Finally, Rizzacasa et al. described the metathesis of a substrate with no phenolic OH and no C15 sidechain with a 77:23 E/Z selectivity [59]. In conclusion, general trends seem to emerge from the previously described RCM on salicylihalamide precursors, but no rational explanation of these effects has been put forward. 2.6.1.6 Oximidines: Use of Relay Ring-closing Metathesis Porco and coworkers employed a M-RCM to synthesize the 12-membered ring of oximidine III [60]. When bis-diene 87 is engaged in a metathesis reaction with catalyst HG [61], whose activity is similar to that of complex G2, diene 88 is produced in only 15% yield, presumably because the ruthenium carbene formed at C10 is unreactive (Scheme 2.35). To circumvent this problem, they used the relay RCM strategy ﬁrst developed by Hoye et al. [62]. The ﬁrst carbene is formed at the terminal oleﬁn of compound 89, and an RCM furnishes the corresponding carbene at C11 (along with cyclopentene), which undergoes a second RCM to produce the desired oximidine core 88 in 71% yield. In this fashion, formation of the stable conjugated carbene at C10 is avoided. OTBS

OTBS

OMe O

OMe O O 15

OTBS

G1

O

CH2Cl2 85

86 (E/Z 20:80)

Scheme 2.34 Effect of the conﬁguration of the C15 stereocenter.

OTBS

j55

56

j 2 Macrocycles by Ring-closure Metathesis OPMB

OPMB TBSO

TBSO

O O

HG

O

10

88 E,Z only

ClCH2CH2Cl 80 °C 71%

O O

10

O

HG

OPMB O

O

CH2Cl2 40 °C 15%

11

87

TBSO

O

Mes N Cl Cl

N Mes Ru

11

O

89

HG Scheme 2.35 Relay RCM in the synthesis of oximidine III.

2.6.2 Radicicol-Type Macrolides: a Promising Family of Anticancer Resorcylides

The tolerance of the RCM reaction to diverse functional groups is illustrated by the synthesis of 14-membered macrocycles monocillin I and radicicol, reported by Danishefsky and coworkers [63]. In the presence of catalyst G2, cyclization of dienic substrates 90 and 91, which comprise an allylic epoxide and a ketone function protected as a dithiane, occurs in 55–60% yields, generating intermediate macrocyclic dienes 92 and 93 of 6E,8Z stereochemistry (Scheme 2.36). It is interesting to

O

H

O

O

O

H

R1O S

G2

S

90 91

1

HO 93

S

O

O

92 93

O

9

O 60%

H

HO

8 7

O

O

SOCl2 Cl

58% OH Monocillin I

55% 6E,8Z 60% 6E,8Z

H

H 6

S

OR2

R2

= Me R = R1 = TBS, R2 = TBDPS

O

H

H

R1O

CH2Cl2 45°C, 5 h

OR2

O

O

O

OH Radicicol

Scheme 2.36 Synthesis of monocillin I and radicicol by M-RCM.

H

2.6 Representative Synthetic Applications

H O

H

O

O H

TBSO S

O H

TBSO S

(5-10 mol%)

OTBDPS 94

42 °C 42 °C 110 °C

19 h 19 h 10 min

S

OTBDPS Yield Dimer 95

Conditions CH2Cl2 Toluene Toluene

S

G2

16% 27% 55%

95

30% 48% -

Scheme 2.37 Problematic RCM in the synthesis of cycloproparadicicol.

note that the absence of protection of the phenolic hydroxyl groups lead to a less efﬁcient RCM reaction. Difﬁculties encountered for the demethylation of the phenolic hydroxyl groups compelled the authors to use the silylated intermediate 93, which can be transformed into monocillin I. Subsequent regioselective chlorination (SOCl2 in Et2O) furnishes radicicol in 58% yield. In order to optimize the antitumoral properties of radicicol, particularly in vivo, the same group synthesized the analogous cycloproparadicicol, where the epoxide function is replaced by a cyclopropane [64]. Submitted to the conditions of the previous RCM reaction (CH2Cl2, 42 C, 19 h), cyclopropyl triene 94 leads to the expected macrolide 95 in only 16% yield, along with 30% of the corresponding 28-membered dimeric macrocycle (Scheme 2.37). After numerous assays, the best conditions tested (toluene, 110 C, 10 min) brought the yield up to 55%. In this case, the balance between thermodynamic and kinetic factors seems decisive for the course of the reaction. The fact that the monomeric product is predominant at elevated temperature indicates that this form is entropically favored. With a view to scaling up the preparation of analogs of radicicol, Yang and Danishefsky developed an ynolide alternative strategy, using the alkyne intermediate 96 [65]. In this route, the macrocycle is constructed without the aromatic nucleus, which is elaborated after formation of the macrocycle by a Diels-Alder reaction with the appropriate cyclohexadiene. However, alkyne 96, when submitted to the RCM conditions, does not furnish macrocycle 97, probably because of the conformational constraints imposed by the triple bond and the transdisubstituted cyclopropane (Scheme 2.38). Deactivation of the catalyst by complexation with the alkyne moiety could also have blocked the metathesis reaction. The problem was circumvented by formation of cobalt complex 98, which possesses an angular conformation (140 angle), allowing the formation of macrocycle 99 with catalyst G2 in 57% yield. The macrocycle was subsequently treated with ceric ammonium nitrate to produce cyclic diene 100, which was converted to cycloproparadicicol by a sequence of reactions involving a Diels-Alder reaction to form the aromatic ring.

j57

58

j 2 Macrocycles by Ring-closure Metathesis H O

H

O

O H

Co2(CO)8, toluene

H

G2 (25 mol%)

O H

O

OTBS 98

OTBS

99 57% H O

H

OTMS

O H

HO

90% OTBS 100

H

TMSO

O

CAN

O

CH2Cl2 (OC)3Co 45 °C, 1.5 h (OC)3Co

H O

OTBS 97

100% H

(OC)3Co (OC)3Co

H

RCM X 96

OTBS

O

O

Cl

O

OH Cycloproparadicicol

Scheme 2.38 Synthesis of cycloproparadicicol: the ynolide strategy.

This ynolide strategy was applied to the synthesis of the macrolide aigialomycin D by Geng and Danishefsky [66]. 2.6.3 Coleophomones: a Versatile Access to this Class of Complex Polycyclic Diterpenes

Coleophomones, isolated by Shionogi Pharmaceutical Co. and later by Merck, are compact fungal diterpenes possessing a strained 11-membered central core. In the ﬁnal strategy adopted by Nicolaou and his group for the synthesis of these challenging molecules, the labile triketone precursor 101 was protected as the mixture of enol ethers 102 and 103 (Scheme 2.39) [67]. After chromatographic separation, these regioisomers were independently submitted to RCM in the presence of G2 to give in each case a single regio- and stereoisomer, the 11-membered rings 104 and 105 possessing an E- and a Z-trisubstituted double bond, respectively. Because of the absence of the more thermodynamically stable cyclopentene adducts, these products were claimed to be the kinetic products of the reaction.

2.6 Representative Synthetic Applications

O

O

O

O

R = p-BrC6H4(CO)

101

RO

CH2N2, Et2O, 0 oC, 1 h

OO

OO

OMe O RO 102 48% (mixture of isomers)

O

RO

103 32% G2 CH2Cl2, 40 oC

G2 CH2Cl2, 40 oC

O

O 8

RO

1

OO 9

OMe

O

O

OMe

105 80%

Z only

RO

104 86%

OMe

E only

Coleophomone B

Coleophomone C

Scheme 2.39 Synthesis of Coleophomones B and C.

2.6.4 RCM in Supramolecular Chemistry

Because of its versatility, high functional-group compatibility, and exceptional catalytic reactivity, the RCM reaction is particularly well suited for supramolecular chemistry applications such as the synthesis of catenanes, rotaxanes, molecular knots, dendrimer-type molecules, etc. [68]. One key strategic point in the construction of such molecular architectures is the application of intermediate use of templates to favor interlocking thermodynamically. Metal-based templation strategy has been successfully applied to the construction of numerous supramolecular ediﬁces. Pioneering work from Grubbs [69] and Sauvage [70] groups, respectively, led to the

j59

60

j 2 Macrocycles by Ring-closure Metathesis O

PF6

O

O

2 PF6 O

O

O O

O O

O O N N

N

N N

N

O

O

N

N

N

N

N

N

O

O

O O

O O

O O

O

O O

106

107

= Cu

= Cu

Figure 2.5 First catenane and molecular knot synthesized by RCM.

formation of the ﬁrst catenane (106) and molecular knot (107) (Figure 2.5) using G1catalyzed RCM of pre-assembled polyoleﬁnic copper complexes. Among the interesting applications of RCM to supramolecular chemistry, the synthesis of molecular gyroscopes containing an internal rotator domain [Fe (CO)3] that can rotate independently of the external stator bearing easily adjustable spokes depending on the diphosphane oleﬁnic chain length has been reported by Gladysz and coworkers [71]. RCM/hydrogenation sequence from the appropriate precursors 108 led to 13-, 15-, and 17-membered gyroscopes 109 (Scheme 2.40). M-RCM supramolecule formation using metal-based templation strategies involving intermediate metal complexes to organize the reacting partners continue to receive much attention since their domain of application can be extended to numerous types of complex products. Conformational constraints such as those dictated by pincer-pyridine complexes clearly favor RCM of terminal oleﬁns 110b and 110c (M¼Pd or Pt) when compared with the unprotected pyridine analog 110a. Final

P OC OC

n

1) G1 (12-19 mol%) CH2Cl2, 40 °C, 7-8 mM 3

OC OC

Fe CO P n

108

2) [ClRh(PPh3)3] (8-15 mol%) H2 (5 atm) 3 toluene, 60-80 °C n = 1-3 79-84%

Scheme 2.40 Formation of molecular gyroscopes by RCM.

P

n

60-81%

Fe CO

n n

P

n

n

109

n

2.6 Representative Synthetic Applications

O

BF4

O

NMe2

O

O 1) G1 (5 mol%)

M N

N 2) NaCl

110

O

NMe2

O

O

O

110a 110b 110c

M

Time

No metal Pd Pt

16 h 5d 2h

Macrocycle 111

111

Isomerized product 50% 0%

20% 50% 100%

Scheme 2.41 Use of pincer-pyridine complexes to favor RCM reaction.

detachment of the pyridine ligands is realized in the presence of NaCl in biphasic systems (Scheme 2.41) [72]. Tripincer template-directed cyclization of a dienic pyridine precursor leads to the formation of the 69-membered ring trimeric heterocycle 112 (Figure 2.6). Decomplexation of the tripincer from the product is realized with NaCl [72]. The synthesis of a [2]rotaxane by M-RCM using a similar pincer-pyridine complex for pre-assembly was also reported by Leigh et al. [73]. Head-to-head a- and b-cyclodextrin homodimers, in which the two primary rims are doubly ligated through alkyl chains, have been efﬁciently synthesized by Sinaÿ and his groups by a CM followed by an RCM reaction sequence [74].

O N

BF4

O Me2 N N Pt O N Me2

NMe2

Pt O Me N 2

Me2N Pt NMe2 O 112

N

O BF4

Figure 2.6 Synthesis of a 69-membered ring by M-RCM.

BF4

j61

62

j 2 Macrocycles by Ring-closure Metathesis 2.7 Conclusion and Perspectives

Ring-closing metathesis has been applied to the synthesis of very diverse macrocycles, ranging from medium to very large rings and containing a wide range of functional groups. These macrocycles can be natural products or complex supramolecular ediﬁces. One key argument that makes a given methodology useful in designing a synthetic process is its predictability, i.e. the possibility of anticipating an expected result from the accumulated knowledge – the state of the art – at a given time. Undoubtedly, today, after no more than ten years of application in macrocyclization applications, the RCM reaction has gained this status and therefore has become an essential tool for achieving the key cyclization steps in macrocycle synthesis. However, some decisive features such as the prevision of the E- or Z-conﬁguration of the resulting double bond have still to be addressed. Several practical solutions to this problem have been proposed and obviously are still going to be found.

2.8 Experimental: Selected Procedures 2.8.1 Synthesis of Compound 3 with Catalyst S1 [12]

Molybdenum catalyst S1 (9 mg, 0.01 mmol) was weighed in a glove box in a 10-mL round-bottom three-necked ﬂask equipped with a reﬂux condenser and a rubber septum. The reaction vessel was subsequently removed from the glove box and maintained under an atmosphere of argon. The catalyst was then dissolved in anhydrous benzene (6.0 mL), and diene 1 (20 mg, 0.04 mmol) was added as a solution in anhydrous benzene (0.60 mL). The reaction ﬂask was immersed in an oil bath that was previously warmed to 50 C, and the mixture was allowed to stir for 14 h. After this period, the reaction vessel was allowed to cool to 22 C. The mixture was stirred in the presence of air (septum was removed) for approximately 2 h. Removal of solvent in vacuo afforded a black solid. Puriﬁcation through silica gel chromatography (10:1 hexanes/ethyl acetate followed by 4:1 hexanes/ethyl acetate) delivered 3 as a white solid (18.1 mg, 90% yield). 2.8.2 Synthesis of Compound 6 with Catalyst G1 [13]

To a solution of 5 (30 mg, 0.114 mmol) in dry degassed CH2Cl2 (25 mL) was added, via syringe, a solution of ruthenium alkylidene catalyst G1 (5 mg, 0.006 mmol) predissolved in CH2Cl2 (6 mL). The resulting purple solution was heated to 45 C for 21 h and then concentrated under reduced pressure to afford an oily brown residue. Puriﬁcation by MPLC [pre-packed silica gel column 2 · 50 cm, eluent: ethyl

2.8 Experimental: Selected Procedures

acetate/n-hexane (1:9)] provided 6 (20.9 mg, 78%). The fractions also provided the cyclic dimer (5 mg, 18%). 2.8.3 Synthesis of Compound 8 with Catalyst G2 [14b]

A stirred solution of diene 7 in anhydrous toluene (0.2 mM) under argon was heated to 110 C (reﬂux), and 5–10 mol% G2 was added to the reﬂuxing solution. Stirring was continued for 15 min, and then the ﬂask was immersed in an ice bath. The reaction mixture was passed through a pad of silica gel, which was rinsed with CH2Cl2 and concentrated, and compound 8 was puriﬁed by ﬂash chromatography. 2.8.4 Synthesis of Compound 16 (R¼SiMe2tBu) with Catalyst G1/Ti(OiPr)4 [17]

The silyl ether 15 (R¼SiMe2tBu) (200 mg, 0.46 mmol) and Ti(OiPr)4 (39 mg, 0.14 mmol, 41 mL, 0.3 equiv) were dissolved in CH2Cl2 (180 mL), and the mixture was reﬂuxed for 1 h. The ruthenium carbene G1 (8 mg, 0.01 mmol, 2 mol%) in CH2Cl2 (2 mL) was then added, and reﬂux was continued for 24 h, after which an additional 4 mg (0.005 mmol, 1 mol%) of G1 in CH2Cl2 (2 mL) was added. After an additional 24 h, the mixture was cooled to ambient temperature and ﬁltered through a short pad of silica gel. The solvent was removed in vacuo, and the residue was puriﬁed by ﬂash chromatography (hexanes/ethyl acetate 100:1) to give the macrolide 16 as a colorless syrup (151 mg, 80%). 2.8.5 Synthesis of Compound 49 by RCAM [38]

Diyne 48 (120 mg, 0.14 mmol) was added to a solution of complex [MoN(Ar)3] (12 mg, 0.02 mmol) in toluene (10 mL) and CH2Cl2 (0.3 mL), and the resulting mixture was stirred at 80 C for 8 h. All volatiles were then evaporated, and the residue was puriﬁed by ﬂash chromatography (pentane/Et2O 20:1) to afford cycloalkyne 49 as a colorless syrup (84 mg, 80%). 2.8.6 Synthesis of Compound 53 with G1 by ene-yne RCM [40a]

To a solution of Grubbs catalyst G1 (943 mg, 1.15 mmol, 0.4 equiv) in CH2Cl2 (900 mL) was added a solution of enyne 51 (2.00 g, 2.86 mmol, 1.0 equiv) in toluene (100 mL). The reaction was purged with ethylene and stirred at 45 C under an atmosphere of ethylene for 40 h. The crude mixture was treated with ethyl vinyl ether (1 mL) and concentrated to a black oil. Puriﬁcation by silica gel chromatography (doped with triethylamine, 5–6% in CH2Cl2/hexane) afforded cyclophane 53 (624 mg, 31%).

j63

64

j 2 Macrocycles by Ring-closure Metathesis Acknowledgments

Financial support was provided by the CNRS and the Ecole Polytechnique. Abbreviations

4CC-IMDA Ac Bn Boc CAN CM Cy DMDO M-RCM MCR Mes MOM P-3CR Piv PMB PNB RCM RCAM ROM TBDPS TBS TES TIPS TPS

4-component condensation/intramolecular Diels-Alder reaction sequence acetyl benzyl tert-butyloxycarbonyl ceric ammonium nitrate cross-metathesis cyclohexyl dimethyldioxirane macrocycle ring-closing metathesis multicomponent reaction mesityl methoxymethyl Passerini 3-component reaction pivaloyl p-methoxybenzyl p-nitrobenzoyl ring-closing metathesis ring-closing alkyne metathesis ring-opening metathesis tert-butyldiphenylsilyl tert-butyldimethylsilyl triethylsilyl triisopropylsilyl triphenylsilyl

References 1 Wessjohann, L.A., Ruijter, E., GarciaRivera, D. and Brandt, W. (2005) Mol. Diversity, 9, 171–186. 2 (a) Balunas, M.J. and Kinghorn, A.D. (2005) Life Sci., 78, 431–441; (b) Butler, M.S. (2004) J. Nat. Prod., 67, 2141–2153; erratum, ibid. 2006, 69, 172; (c) Shu, Y.-Z. (1998) J. Nat. Prod., 61, 1053–1071. 3 (a) For a comprehensive review on metathesis, see: Grubbs, R.H. (ed.)

(2003) Handbook of Metathesis, WileyVCH, Weinheim; (b) for a review of M-RCM in natural product synthesis, see:Gradillas, A. and Pérez-Castells, J. (2006) Angew. Chem., 118, 6232– 6247; Angew. Chem. Int. Ed., 45, 6086–6101. 4 (a) Schrock, R.R., Murdzek, J.S., Bazan, G.C., Robbins, J., DiMare, M. and O’Regan, M. (1990) J. Am. Chem. Soc., 112, 3875–3886; (b) For a review:

References

5

6

7

8

9 10

11

12

13

14

15 16

Schrock, R.R. and Hoveyda, A.H. (2003) Angew. Chem., 115, 4740–4782; (c) Schrock, R.R. and Hoveyda, A.H. (2003) Angew. Chem. Int. Ed., 42, 4592–4633. Nguyen, S.T., Grubbs, R.H. and Ziller, J. W. (1993) J. Am. Chem. Soc., 115, 9858– 9859. (a) Schwab, P., France, M.B., Ziller, J.W. and Grubbs, R.H. (1995) Angew. Chem., 107, 2179–2181; Angew. Chem. Int. Ed., 34, 2039–2041; (b) Schwab, P., Grubbs, R.H. and Ziller, J.W. (1996) J. Am. Chem. Soc., 118, 100–110. Scholl, M., Trnka, T.M., Morgan, J.P. and Grubbs, R.H. (1999) Tetrahedron Lett., 40, 2247–2250. (a) Huang, J., Stevens, E.D., Nolan, S.P. and Petersen, J.L. (1999) J. Am. Chem. Soc., 121, 2674–2678; (b) Jafarpour, L., Hillier, A.C. and Nolan, S.P. (2002) Organometallics, 21, 442–444. Scholl, M., Ding, S., Lee, C.W. and Grubbs, R.H. (1999) Org. Lett., 1, 953–956. (a) Astruc, D. (2005) New J. Chem., 29, 42–56; ibid. (2006), 30, 1848–1852; (b) Katz, T.J. (2006) New J. Chem., 30, 1844–1847. Nicolaou, K.C., Bulger, P.G. and Sarlah, D. (2005) Angew. Chem., 117, 4564–4601; Angew. Chem. Int. Ed., 44, 4490–4527. Xu, Z., Johannes, C.W., Houri, A.F., La, D.S., Cogan, D.A., Hoﬁlena, G.E. and Hoveyda, A.H. (1997) J. Am. Chem. Soc., 119, 10302–10316. Kamat, V.P., Hagiwara, H., Katsumi, T., Hoshi, T., Suzuki, T. and Ando, M. (2000) Tetrahedron, 56, 4397–4403. (a) Yamamoto, K., Biswas, K., Gaul, C. and Danishefsky, S.J. (2003) Tetrahedron Lett., 44, 3297–3299; (b) See also: Biswas, K., Lin, H., Njardarson, J.T., Chappell, M.D., Chou, T.-C., Guan, Y., Tong, W.P., He, L., Horwitz, S.B. and Danishefsky, S.J. (2002) J. Am. Chem. Soc., 124, 9825–9832. Fürstner, A. and Langemann, K. (1997) Synthesis, 792–803. Goldring, W.P.D., Hodder, A.S. and Weiler, L. (1998) Tetrahedron Lett., 39, 4955–4958.

17 Fürstner, A. and Langemann, K. (1997) J. Am. Chem. Soc., 119, 9130–9136. 18 (a) Baltrusch, A.W. and Bracher, F. (2002) Synlett, 1724–1726; (b) for yet another example of the use of Ti(OiPr)4 in M-RCM, see: Eh, M. (2003) Synthesis, 365–370. 19 Marsella, M.J., Maynard, H.D. and Grubbs, R.H. (1997) Angew. Chem., 109, 1147–1150; Angew. Chem. Int. Ed., 36, 1101–1103. 20 Maleczka, R.E., Jr, Terrell, L.R., Geng, F. and Ward III, J.S., (2002) Org. Lett., 4, 2841–2844. 21 Hoye, T.R. and Zhao, H. (1999) Org. Lett., 1, 1123–1125. 22 Evano, G., Schaus, J.V. and Panek, J.S. (2004) Org. Lett., 6, 525–528. 23 Sello, J.K., Andreana, P.R., Lee, D. and Schreiber, S.L. (2003) Org. Lett., 5, 4125– 4127. 24 Banwell, M.G. and McRae, K.J. (2000) Org. Lett., 2, 3583–3586. 25 For a review on the E/Z selectivity in MRCM reactions, see: Prunet, J. (2003) Angew. Chem., 115, 2932–2936; Angew. Chem. Int. Ed., 42, 2826–2830; erratum, Angew. Chem. (2003), 115, 3444; (2003) Angew. Chem. Int. Ed., 42, 3322. 26 Lee, C.W. and Grubbs, R.H. (2000) Org. Lett., 2, 2145–2147; erratum, ibid. 2000, 2, 2558. 27 (a) For the ﬁrst example of such a control in RCM, see: Bourgeois, D., Pancrazi, A., Ricard, L. and and Prunet, J. (2000) Angew. Chem., 112, 741–744; (2000) Angew. Chem. Int. Ed., 39, 725–728; (b) Bourgeois, D., Mahuteau, J., Pancrazi, A., Nolan, S.P. and Prunet, J. (2000) Synthesis, 869–882. 28 Fürstner, A., Radkowski, K., Wirtz, C., Goddard, R., Lehmann, C.W. and Mynott, R. (2002) J. Am. Chem. Soc., 124, 7061–7069. 29 Garcia-Fortanet, J., Murga, J., Falomir, E., Carda, M. and Marco, J.A. (2005) J. Org. Chem., 70, 9822–9827. 30 Arisawa, M., Kato, C., Kaneko, H., Nishida, A. and Nakagawa, M. (2000)

j65

66

j 2 Macrocycles by Ring-closure Metathesis 31 32

33

34

35

36

37

38

39

J. Chem. Soc., Perkin Trans. 1, 1873–1876. Nakashima, K., Ito, R., Sono, M. and Tori, M. (2000) Heterocycles, 53, 301–314. Jasper, C., Wittenberg, R., Quitschalle, M., Jakupovic, J. and Kirschning, A. (2005) Org. Lett., 7, 479–482 erratum, ibid. 2005, 7, 1885. Nicolaou, K.C., He, Y., Vourloumis, D., Vallberg, H., Roschangar, F., Sarabia, F., Ninkovic, S., Yang, Z. and Trujillo, J.I. (1997) J. Am. Chem. Soc., 119, 7960– 7973. Meng, D., Su, D.-S., Balog, A., Bertinato, P., Sorensen, E.J., Danishefsky, S.J., Zheng, Y.-H., Chou, T.-C., He, L. and Horwitz, S.B. (1997) J. Am. Chem. Soc., 119, 2733–2734. Nicolaou, K.C., Ninkovic, S., Sarabia, F., Vourloumis, D., He, Y., Vallberg, H., Finlay, M.R.V. and Yang, Z. (1997) J. Am. Chem. Soc., 119, 7974–7991. Fürstner, A. and Seidel, G. (1998) Angew. Chem., 110, 1758–1760; Angew. Chem. Int. Ed., 37, 1734–1736. (a) Other syntheses by Fürstner et al. using this strategy: prostaglandin E21,15-lactone: Fürstner, A., Grela, K., Mathes, C. and Lehmann, C.W. (2000) J. Am. Chem. Soc., 122, 11799–11805; (b) sophorolipid lactone: Fürstner, A., Radkowski, K., Grabowski, J., Wirtz, C. and Mynott, R. (2000) J. Org. Chem., 65, 8758–8762; (c) ambrettolide and yuzu lactone: Fürstner, A., Guth, O., Rumbo, A. and Seidel, G. (1999) J. Am. Chem. Soc., 121, 11108–11113; (d) motuporamine C: Fürstner, A. and Rumbo, A. (2000) J. Org. Chem., 65, 2608–2611; (e) civetone: Fürstner, A. and Seidel, G. (2000) J. Organomet. Chem., 606, 75–78; (f) turrianes: Fürstner, A., Stelzer, F., Rumbo, A. and Krause, H. (2002) Chem. Eur. J., 8, 1856–1871. Fürstner, A., Mathes, C. and Lehmann, C.W. (2001) Chem. Eur. J., 7, 5299– 5317. Fürstner, A. and Radkowski, K. (2002) Chem. Commun., 2182–2183.

40 (a) Layton, M.E., Morales, C.A. and Shair, M.D. (2002) J. Am. Chem. Soc., 124, 773–775; (b) Morales, C.A., Layton, M.E. and Shair, M.D. (2004) Proc. Natl. Acad. Sci. USA, 101, 12036–12041. 41 Smith, A.B., III, Adams, C.M., Kozmin, S.A. and Paone, D.V. (2001) J. Am. Chem. Soc., 123, 5925–5937. 42 Lazarova, T.I., Binet, S.M., Vo, N.H., Chen, J.S., Phan, L.T. and Or, Y.S. (2003) Org. Lett., 5, 443–445. 43 Vassilikogiannakis, G., Margaros, I. and Toﬁ, M. (2004) Org. Lett., 6, 205–208. 44 Lee, C.W., Choi, T.-L. and Grubbs, R.H. (2002) J. Am. Chem. Soc., 124, 3224– 3225. See also Ref. 3, Vol. 2, p. 238 for a review on ring-expansion metathesis. 45 Louie, J., Bielawski, C.W. and Grubbs, R.H. (2001) J. Am. Chem. Soc., 123, 11312–11313. 46 (a) Beck, B., Larbig, G., Mejat, B., Magnin-Lachaux, M., Picard, A., Herdtweck, E. and Dömling, A. (2003) Org. Lett., 5, 1047–1050; (b) Dömling, A. (2006) Chem. Rev., 106, 17–89. 47 Van de Weghe, P. and Eustache, J. (2005) Curr. Top. Med. Chem., 5, 1495–1519. See also Ref. 3a, Vol. 2, p. 353 for epothilones and p. 54 for peptidomimetics. 48 Yet, L. (2003) Chem. Rev., 103, 4283–4306. 49 Wu, Y., Liao, X., Wang, R., Xie, X.-S. and De Brabander, J.K. (2002) J. Am. Chem. Soc., 124, 3245–3253. 50 Herb, C., Dettner, F. and Maier, M.E. (2005) Eur. J. Org. Chem., 728–739. 51 Haack, T., Haack, K.L., Diederich, W.E., Blackman, B., Roy, S., Pusuluri, S. and Georg, G.I. (2005) J. Org. Chem., 70, 7592–7604. 52 Georg, G.I., Ahn, Y.M., Blackman, B., Farokhi, F., Flaherty, P.T., Mossman, C. J., Roy, S. and Yang, K.L. (2001) Chem. Commun., 255–256. 53 Fürstner, A., Dierkes, T., Thiel, O.R. and Blanda, G. (2001) Chem. Eur. J., 7, 5286– 5298. 54 Labrecque, D., Charron, S., Rej, R., Blais, C. and Lamothe, S. (2001) Tetrahedron

References

55 56 57 58 59

60

61

62

63

64

Lett., 42, 2645–2648. The metathesis reaction is performed on the enantiomer of 83 in this case. Smith, A.B., III and Zheng, J. (2002) Tetrahedron, 58, 6455–6471. Yadav, J.S. and Srihari, P. (2004) Tetrahedron: Asymmetry, 15, 81–89. Snider, B.B. and Song, F. (2001) Org. Lett., 3, 1817–1820. Herb, C., Bayer, A. and Maier, M.E. (2004) Chem. Eur. J., 10, 5649–5660. Feutrill, J.T., Holloway, G.A., Hilli, F., Hügel, H.M. and Rizzacasa, M.A. (2000) Tetrahedron Lett., 41, 8569–8572. Wang, X., Bowman, E.J., Bowman, B.J. and Porco, J.A., Jr (2004) Angew. Chem., 116, 3685–3689; Angew. Chem. Int. Ed., 43, 3601–3605. Garber, S.B., Kingsbury, J.S., Gray, B.L. and Hoveyda, A.H. (2000) J. Am. Chem. Soc., 122, 8168–8179; erratum, ibid. 2001, 123, 3186. (a) Hoye, T.R., Jeffrey, C.S., Tennakoon, M.A., Wang, J. andZhao, H. (2004) J. Am. Chem. Soc., 126, 10210–10211; (b) for a review on relay RCM, see: Wallace, D.J. (2005) Angew. Chem., 117, 1946–1949; Angew. Chem. Int. Ed., 44, 1912–1915. Garbaccio, R.M., Stachel, S.J., Baeschlin, D.K. and Danishefsky, S.J. (2001) J. Am. Chem. Soc., 123, 10903–10908. Yamamoto, K., Garbaccio, R.M., Stachel, S.J., Solit, D.B., Chiosis, G., Rosen, N.

65 66 67

68

69

70

71

72

73

74

and Danishefsky, S.J. (2003) Angew. Chem., 115, 1318–1322; Angew. Chem. Int. Ed., 42, 1280–1284. Yang, Z.-Q. and Danishefsky, S.J. (2003) J. Am. Chem. Soc., 125, 9602–9603. Geng, X. and Danishefsky, S.J. (2004) Org. Lett., 6, 413–416. Nicolaou, K.C., Montagnon, T., Vassilikogiannakis, G. and Mathison, C. J.N. (2005) J. Am. Chem. Soc., 127, 8872– 8888. For reviews on RCM in supramolecular chemistry, see Ref. 3a, Vol. 2, pp. 96 and 410. Weck, M., Mohr, B., Sauvage, J.-P. and Grubbs, R.H. (1999) J. Org. Chem., 64, 5463–5471. Dietrich-Buchecker, C., Rapenne, G. and Sauvage, J.-P. (1997) Chem. Commun., 2053–2054. Shima, T., Hampel, F. and Gladysz, J.A. (2004) Angew. Chem., 116, 5653–5656; Angew. Chem. Int. Ed., 43, 5537–5540. Chase, P.A., Lutz, M., Spek, A.L., van Klink, G.P.M. and van Koten, G. (2006) J. Mol. Catal. A, 254, 2–19. Fuller, A.-M., Leigh, D.A., Lusby, P.J., Oswald, I.D.H., Parsons, S. and Walker, D.B. (2004) Angew. Chem., 116, 4004– 4008; Angew. Chem. Int. Ed., 43, 3914– 3918. Lecourt, T., Mallet, J.-M. and Sinaÿ, P. (2003) Eur. J. Org. Chem., 4553–4560.

j67

j69

3 Supramolecular Macrocycle Synthesis by H-bonding Assembly Pablo Ballester and Javier de Mendoza

3.1 Introduction

A supramolecular macrocycle can be considered as a non-covalent cyclic oligomer formed by n monomers of the same or different chemical type. The formation of such architecture requires the establishment, of at least n 1 intermolecular interactions and one intramolecular interaction. In fact, the intramolecular interaction appears as the necessary requisite to produce large, complex but discrete and deﬁned oligomers in solution [1]. From this deﬁnition it can be inferred that the molecular components of supramolecular macrocycles have to be multitopic. For example, the self-assembly of a simple two-dimensional cyclic structure from a single component demands a difunctional monomer. Because of the non-covalent character of supramolecular assembly, the products are equilibrating structures. Therefore, the selfassembly of a cyclic oligomer will always be in competition with the self-assembly of linear oligomers and possibly other cyclic aggregates. This chapter deals with formation of cyclic 2D and 3D structures in solution by self-assembly of two or more components using hydrogen bonds as the major interactions. Obviously, the formation of a deﬁned supramolecular aggregate stabilized by non-covalent forces is a thermodynamically driven process which reﬂects a balance between enthalpy and entropy. Consequently, the product of a non-covalent macrocyclic synthesis must be evaluated and predicted in terms of thermodynamic minima in an equilibrium mixture. In chloroform solution, the enthalpy of an isolated neutral hydrogen bond ranges from 1 to 7 kcal mol1. Because of the enthalpy-entropy compensation effect, the entropic cost for the formation of a bimolecular aggregate stabilized by hydrogen bonds could also be quite variable. In addition, use of multiple hydrogen bonds can yield a stronger interaction than that provided by the sum of individual donor-acceptor partners, because of secondary attractive interactions. Thus, the free enthalpy (DG) of binding of neutral hydrogen-bonded complexes can be estimated and used in a predictable manner for a number of systems [2]. In chloroform, for instance, and

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

70

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly based on the relative hydrogen-bonding afﬁnities of imides and lactams, it is possible to assign a free enthalpy change of 1.5 to 1.9 kcal mol1 for the establishment of a primary NH O¼C hydrogen bond and a cost of 0.6–0.7 kcal mol1 for each secondary repulsive interaction between spectator oxygen atoms [3]. However, a secondary interaction of attractive nature will contribute to the stability of the complex to a similar extent. This simple back-of-the-envelope approximation is not quantitatively perfect, but it does reﬂect some fundamental variations in systems stabilized by hydrogen-bonded arrays [4]. The theoretical treatment for the self-assembly of a molecular entity into rings or cages occurring under thermodynamic control is based on two fundamental physicochemical quantities: the effective molarity (EM) of the intramolecular interaction of the cyclic n-mer and the value of the stability constant of the intermolecular interaction (Kinter) [5]. The following formula has been derived by Ercolani and allows calculation of the minimum value of the product KinterEM to obtain a virtually complete self-assembled macrocycle at certain monomer concentration: K inter EM

Q1

1=ðBNþ1Þ N2 185 B

ð1Þ

where N is the number of molecular units involved in the supramolecular structure of the macrocycle (minimum 2) and B the overall number of intermolecular bonds holding these units together. Thus, it turns out that for N ¼ 2 and B ¼ 2 the complete self-assembly requires a driving force KinterEM 370. The upper limit of the EM of a supramolecular macrocycle stabilized by intermolecular interactions has been estimated on the order of 10 M [2]. Consequently, because of the low gain of free enthalpy obtained in the formation of a single intermolecular hydrogen bond (1.5 kcal mol1, Kinter ¼ 12 M1), the efﬁcient and complete self-assembly of a supramolecular macrocycle using neutral hydrogen bonds requires the use of subunits containing multiple hydrogen bonds. As the equation points out, an increase in the number of bonds (B) holding together the pieces of an assembly decreases the required driving force of the assembly. The increase of the Kinter favors the fulﬁllment of the above-mentioned relationship. In 1992, Zimmerman and Duerr reported the use of the hydrogen bond-mediated cyclic aggregation strategy in the construction of cyclic trimer 1 from monomer 2 as opposed to dimer 3 or the open trimer 4 (Scheme 3.1) [6]. The cyclic structure of the aggregate and its corresponding trimerization constant (Ktrimer ¼ 20000 M2) were estimated from the Saunders-Hyne analysis of the 1 H NMR dilution shifts of 1 in 10% (CD3)2SO/CDCl3. Model systems 5 and 6 did not self-associate or bind to one another to a signiﬁcant extent in the same solvent mixture. Consequently, the authors assigned values of K7 < 3 M1 for the formation of the complex 7 and Kdimer < 3 M1 for self-aggregation (Scheme 3.2). The value of K7 3 M1 could be used as a reference for the equilibrium constant of the intermolecular interaction (Kinter) of the cyclic trimer 1. Consequently, since Ktrimer ¼ EMK 3inter , from Ktrimer ¼ 20000 M2 an EM 740 M can be calculated.

N

H

N

N

H

O

O

1

O

H

N

O

N H

O H N H

N O

Scheme 3.1 Self-assembly of a cyclic trimer and open oligomers.

O

H

2

N H

O

H

N O

N H O

O H N

O N

H

H

N O

n = 0; 3 n = 1; 4

N H

O n

3.1 Introduction

j71

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly Bu

Bu Bu +

N H 5

H

Bu

N

O

O

H

N H

O

N

O 7

6

Scheme 3.2 Heterodimer formation.

According to the equation derived by Ercolani, this EM value is compatible with the complete assembly of the trimer (KinterEM ¼ 3 740 185 n ¼ 185 3) at certain concentrations of the monomer. The calculated speciation proﬁles during the range of concentration of the dilution experiment are shown in Figure 3.1. The binding model used for the simulation assumes the formation of trimer 1 and two other species: dimer 3 (n ¼ 0) and an open trimer 4 (n ¼ 1). The stability constants for these aggregates were estimated as K3 ¼ K7 and K 4 ¼ K 27 . As shown in Figure 3.1, the molar fraction that could be reached for the macrocycle in the experimental range of monomer concentration is about 85%. A quantitative formation of the macrocycle (>98%) would require working at a monomer concentration close to 2 M. The takehome message is that even a supramolecular macrocyclic system having a high stability constant and fulﬁlling Ercolani's equation may require a high concentration of the monomer to be formed quantitatively. If we assume that the EM value has a limit close to a two-digit number, then the only practical way to reduce the monomer concentration and to achieve a quantitative cyclic assembly stabilized by hydrogen bonds is to increase the value of the intermolecular interaction (Kinter). For the case at hand, maintaining the same EM value and increasing the Kinter just one order of magnitude, i.e. from 3 M1 to 30 M1, the quantitative formation of the cyclic trimer would require a monomer concentration of just 5 102 M, that is, 40-fold smaller than before. The most intuitive way to increase the stability of an intermolecular interaction established through hydrogen bonds is to increase the number of hydrogen bonds formed or, alternatively, to work in less competitive solvents. The use of hydrogen bond-mediated cyclic aggregation for the effective synthesis of macrocycles requires 100 80 % of species

72

Monomer Dimer Open Trimer Cyclic Trimer

60 40 20 0 –4

–3

–2

–1

log [Monomer concentration] Figure 3.1 Calculated speciation proﬁle during Zimmermans dilution experiment [6].

3.1 Introduction

N H N H

N H

O 8

O O

N H

N H 9

O

O H N

O

H N

O

O

N H

O

N H

N H

O

H N

j73

O

O H N

O

N H +

N H O

O H

N

Scheme 3.3 Self-assembly behavior of self-complementary (8) and non-self-complementary (9) dipyridone modules.

the formation of multiple arrays of intermolecular hydrogen bonds between its molecular components. Four years before Zimmerman and Duerrs publication, Ducharme and Wuest [7] showed how the design of intermolecular hydrogen bonding arrays between ditopic monomers could be used to control molecular aggregation. The self-complementary asymmetric dipyridone 8 forms almost exclusively and quantitatively a dimer (Scheme 3.3) in chloroform at monomer concentrations as low as 3.6 · 104 M. The authors estimated that the stability constant for the dimer was higher than 5 · 104 M1. They also calculated a Kdimer ¼ 95 M1 for the dimerization of 2-pyridone under the same conditions. Taking this value as reference for the Kinter an EM ¼ 6.48 M can be calculated for Wuests system. The dimer is stabilized by the formation of two hydrogen bonds at each of the sticky ends. Because of geometric constraints, the symmetric dipyridone isomer 9 is only capable of establishing two intermolecular hydrogen bonds when forming a cyclic or an open dimer. Assigning an EM = 6.48 M as calculated for 8 and a Kinter ¼ 12 M1 (in chloroform DG 1.5 kcal/mol for an NH O¼C hydrogen bond), Ercolanis equation (KinterEM ¼ 77 < 370) predicts the cyclic dimer not to be formed completely at any monomer concentration. From the calculated dimerization value for 2-pyridone (Kdimer ¼ 95 M1) we estimate the stability constant for the open dimer as Kopendimer 4 · 95 ¼ 380 M1. The infrared spectrum of dipyridone 9 indicates only 20% of dimerization at a monomer concentration of 1.1 · 103 M. This is a concentration value three times higher than that required for the quantitative formation of the cyclic dimer of 8. In the solid state it was observed that 9 adopts a polymeric motif of the open dimer. With the above simple examples we wish to emphasize that the synthetic use of hydrogen bonding assemblies in the effective formation of supramolecular macrocycles relies on high EM values and on the establishment of stable intermolecular arrays of hydrogen bonds capable of affording a high enthalpic gain to compensate for the entropic cost of the assembly.

H N

O

74

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly H O

H

11

R1 N

N

N

H

R2

H

H

R N

N H

N

H H

H

N

N N

N

R N

N

1

R

H N

H

H N

N

R1

N

N

H

O N

N

O

H

H H

O

O

CA3. M3

N

N H

H

N H

H O

N H

R1 N

N N

N

N

O

1

1

O

O

H

R1 N

O

H O

N 10 R2

O

R2

H

N

N

N

R1 N

N

N H

H N

H N

R1

N

H H

N R2

R2 O

N N

N

H O

N

N

R1

O

R1 = R2 =

Scheme 3.4 Tape vs. hexagonal rosette formation in Whitesides melamine-cyanurate aggregates [10].

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

The cyclic aggregate CA3M3 rosette system introduced by Seto and Whitesides in1993 elegantly illustrates most of the principles discussed in the Introduction (Scheme 3.4) [8]. This system utilizes the hydrogen-bonded pattern of the cyanuric acid CA (10) - melamine M (11) lattice [9,10]. The components of the CA-M dimer are two heterocyclic units acting as difunctional monomers ideally designed for the construction of linear and cyclic motifs. The maximum enthalpy for the formation of the cyclic aggregate in the simple CA3M3 motif corresponds to 18 hydrogen bonds. In chloroform, the authors estimated an enthalpic gain upon formation of such a hydrogen-bonded network of ca. 24 kcal mol1 (1.3 kcal mol1 for each hydrogen bond).1) However, formation of a cyclic aggregate by interaction of six molecules 1) Our simple back-of-the-envelope approxima-

tion assigns a Kinter ¼ 100 M-1 (three primary interactions and four repulsive secondary interactions). Assuming an EM ¼ 10, it turns out that KinterEM ¼ 1000. Ercolanis minimum value to obtain a virtually complete self-assembled macrocycle for such a system (considering each triply hydrogen-bonded interaction as one intermolecular interaction) turns out to be 1110. The similarity of these two values indicates that the complete self-assembly of the macrocycle

requires high concentrations of the monomers and reﬂects a low thermodynamic stability. In fact, without preorganization via the hub and the spokes, CA and M derivatives associate in solution to form hydrogen-bonded oligomers that exchange rapidly at room temperature. Only at 60 C do the NH protons that are broad at room temperature become sharp and shift downﬁeld, indicating the existence of a predominant geometry, probably the cyclic hexamer.

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

involves a high entropic loss. The translational and rotational entropic costs of bringing together two molecules to form a complex stabilized by two hydrogen bonds has been estimated to be about 3–4.5 kcal mol1 in chloroform [11]. Therefore, the sum of all unfavorable entropy contributions (18–27 kcal mol1) [12] could hardly be balanced by the overall enthalpic gain (24 kcal mol1), and the rosette should not form, unless something is done to reduce the entropy penalty. Preorganization and peripheral crowding were employed by Whitesides and coworkers in their rosettes as successful strategies [9]. Preorganization lies at the foundation of molecular recognition [13], and the concept directly points to a reduction of the entropic cost by conformational control, usually introduced during the synthesis of the molecular components that self-assemble. This means that in a simple bimolecular aggregate some molecular components are introduced during the synthesis of the interacting partners that shape their conformations close to those found in the complex. For a rosette with multiple contacts, however, the concept must not only be applied to the conformational organization of the melamine subunits close to that achieved on complexation but also to reduction of the number of degrees of rotational and translational freedom that are lost upon aggregation. The covalent attachment of three components of the rosette (usually the melamines for synthetic reasons) to a common hub by appropriate connectors restricts the rotations into a conformation close to that observed in the CA3M3 assembly. It also diminishes from 3 to 1 the number of melamine molecules that will lose translational and rotational entropy after self-assembly. The unfavorable conformational, translational, and rotational entropic penalties associated with the formation of the aggregate have been partially paid off by the new tripodal melamine component. Although less effective, intermolecular steric hindrance between substituents is another approach to achieving selection of the cyclic CA3M3 rosette in preference to competitive linear motifs in both the solid and solution states (Figure 3.2) [14]. Choosing the substituents on the melamine and cyanuric acid groups in such a

CA3.M3

Figure 3.2 Peripheral crowding as a strategy to prevent linear aggregates.

j75

76

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly O

O

O

N H

O

N

H2N

13

O H N

12 N N

O H N

Br

N N H

H2N

N N

N N H

Figure 3.3 Schematic representation of tripodal melamines with linkers (12 and 13) to promote formation of rosettes [9].

way that non-cyclic structures become energetically less favored (for steric reasons) promotes formation of the cyclic assembly. This is another example of preorganization which relies on enthalpic rather than entropic control. Complexes with higher enthalpies having a similar entropic component will be formed to a lesser extent. Furthermore, the cyclic oligomer may also be favored relative to open oligomers because the interaction that closes the cycle is intramolecular and can beneﬁt from the chelate effect (effective molarity). Inmanycases,however,thepreorganizationofareceptormakesitdifﬁculttoseparate the effects of entropy and enthalpy upon binding. Conformational preorganization of a receptormay inducerepulsiveenthalpicinteractions between nearby functionalgroups or substituents (i.e. melamine residues of the tripodal construct). The elimination or reduction of this enthalpic destabilization in the free preorganized unit can be the main driving force for complex formation instead of simple preorganization. The observed difference in stability of the rosette complexes formed with tripodal melamine subunits having the rigid (12) or the more ﬂexible (13) is consistent with the entropy gain due to restricted conformational ﬂexibility of three single bonds (Figure 3.3). Whitesides combined both strategies to create a wide range of self-assembled oligocyclic aggregates containing mono- and bis-rosette motifs. He undertook a complete, detailed and accurate experimental program to study and characterize the structures and to measure their relative stabilities [10]. A related thermodynamic analysis was carried out by Hunter and coworkers using the strict self-assembly of cyclic porphyrin oligomers [15]. One of the most important conclusions of these studies is that the DG value is not an appropriate parameter to compare the relative stability of cyclic multiparticle aggregates containing different numbers of molecular components. This is because the effect of concentration (and, therefore, entropy) is not indicated in DG (which assumes that aggregates and components are at 1 M). Let us compare the CA-M dimer to the cyclic hexameric aggregates with estimated KCA-M ¼ Kinter ¼ 102 M1 and Kcyclic(CA-M)6 ¼ (Kinter)6 EM ¼ (102)6 · 10 ¼ 1013 M5 2). Initially, it might be wrongly concluded that the hexamer is always more stable, as it has the highest K value and, consequently, a more negative DG .

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds Table 3.1 Stability indices for hydrogen-bonded rosettes.

Na HBb HB/(N 1)c wDMSOd

6 18 3.6 0.04

9 36 4.5 0.23

4 18 6 0.4

4 36 12 0.7

a

Number of molecules; Number of hydrogen bonds; c This ratio reﬂects the competition contributions of enthalpy and entropy; d Mole fraction of DMSO in chloroform where half of the aggregate has decomposed into components. b

A simple speciation simulation, however, shows that at [CA]o ¼ [M]o ¼ 1 mM the rosette CA3M3 is almost not formed at all (only around 1.5% of the components are self-assembled this way), whereas the dimer CA-M consumes more than 8% of the components. At this concentration ca. 90% of the components are free. A ten-fold increase in concentration, [CA]o ¼ [M]o ¼ 10 mM, highly favors the conversion of monomers to assemblies and changes the ratio of the formed aggregates. Up to 70% of monomers are now assembled as hexamers and only 6% as dimers.3) The percentage of conversion of monomers to assemblies depends strongly on the number of monomers N and the concentration used for the self-assembly experiment. It is worthy of note that Ka or DG do not correlate in a simple way to the conversion of the monomeric components into the full assembly. To overcome this limitation, Whitesides and coworkers introduced three indices to estimate the relative stability of hydrogen-bonded cyclic assemblies based on N (number of molecules) and HB (number of hydrogen bonds). Conceptually, the ratio HB/(N 1) reﬂects the competition contributions of enthalpy (DH is proportional to HB) and entropy (DS is proportional to N). The authors observed that the stability measured experimentally from studies of de-aggregation in DMSO correlates with the value of the ratio HB/(N 1). Table 3.1 shows the corresponding values for selected aggregates [9,10]. The most stable aggregate of Whitesides rosettes family shown in Table 3.1 uses a combination of preorganization (sideways connection of the cyanurate subunits) and peripheral crowding of the melamine subunits for the preferential formation of 2) For simplicity the symmetry correction factors

are not considered.

3) For KCA-M ¼ 103 M-1 and Kcyclic(CA-M)6 ¼ 1013 M-5

at [CA]o ¼ [M]o ¼ 10 mM, 45% of monomers are assembled as dimers and only 32% as hexamers.

j77

78

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly NH2

N HN

NH2

N

N

O

mirror

N

N

NHHN

N

O

O

O

NH

P

M D3

14 O H

N

N

O

H O C3h

15

mirror NH2

N HN

C5

N

O

NH2

N

N

N

NHHN

N

O

O

O

M

N NH

HN

NH2

H 2N

N

N

N

NHHN

O

O

staggered

O

N N

NH

O

P

Figure 3.4 Origin of chirality in Reinhoudts double rosettes.

the rosette. Reindhoudt and coworkers showed that pinched-cone calix[4]arenes serve as excellent linkers for double rosette assemblies (HB/(N 1) ¼ 4.5) [16]. The diametrical substitution of a calix[4]arene with two melamine fragments at the upper rim (14) provides sufﬁcient preorganization for the efﬁcient self-assembly with two equivalents of cyanurate or barbiturate (15) into double rosettes with improved thermodynamic stability as compared to that of single rosettes (Figure 3.4). These assemblies can exist in three different conformations with D3, C3h, or C5 symmetries. In the D3 conformation, the two melamine fragments of the calix[4] arene component adopt an antiparallel (staggered) orientation, which renders the assembly chiral. The chirality arises from the fact that the two melamines can adopt either a clockwise (P) or counter-clockwise (M) orientation (Figure 3.4). Once formed, the double rosette is kinetically stable, and both enantiomers do not interconvert in organic solvents, as this would require an almost complete

Scheme 3.5 Diastereoselective formation of double rosettes.

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

j79

80

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly H

R N

O R D A D A A D DDA AA D A A D D D A

A D A D D A DA A ADD D A D A A D

H N

R

N H

N H

O

N

R

O H N

N N R

N

N

N

N N

O

H

H

N

H H O H

H O H H

N

N

H O

H

N O

N H

R N

N

N R

H

H

H N

O 16 N

N

N

N N

N H O

R

N

O

H N

H N

R

N N

N

R

N

H

R

O N R

H

Figure 3.5 Proposed self-assembly of the self-complementary molecule 16 (R ¼ C8H17) into a supramolecular macrocycle; schematic (left, A ¼ acceptor, D ¼ donor) and structural (right) representation.

dissociation of the aggregate, a too energetically costly process [17]. Introduction of one chiral component into the mixture, such as a chiral barbiturate or a chiral melamine, results in the remarkable formation of a preferred diastereoisomer in high stereoselectivity (Scheme 3.5) [18]. Moreover, once chirality has been induced via an enantiomerically pure barbiturate, the chiral component can be replaced in the rosette by an achiral cyanurate without destroying the P or M self-assembly (chiral memory concept) [17,19]. Unlike Whitesides rosettes, based on symmetric ADA-DAD (A ¼ acceptor, D donor) triple hydrogen-bonding motifs, introduction of DAA and ADD edges within a 60 angle in a planar molecule would also result in hexameric macrocycles, without competition with linear aggregates (Figure 3.5). Such Janus type molecules (16), hydrogen-bond-programmed to self-assemble into cyclic arrays, were ﬁrst reported by Lehn and coworkers in 1994 [20]. Aggregation was studied by NMR spectroscopy, and vapor pressure osmometry (VPO) and gel permeation chromatography (GPC) measurements reasonably agreed with the formation of hexameric rosettes. Similar molecules (17) were described by Mascal as G-C DNA base hybrids [21]. In this case, the hexameric macrocycle was beautifully evidenced in the solid state by an X-ray structure (Figure 3.6). The limited solubility in organic solvents of most monomers endowed with hydrogen-bonding edges often prevents the achievement of the high concentrations (high EM values) required for complete cyclization. Therefore, efforts have concentrated on enhancing the intermolecular contacts to raise the enthalpy gain of the interaction. The two strategies most frequently applied are: (a) to add new contacts at

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

Figure 3.6 Mascals G-C DNA base hybrids in the solid state. Reproduced with permission of Ref. [21b].

Figure 3.7 A rosette with peripheral stabilization by metal coordination [22].

j81

82

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly OC12H25 OC12H25 OC12H25 O R

N

O H

H N

H N R H25C12O H25C12O

N H O

N

H N

O N

H25C12O O

N

H

H

H

H

N R

H

N

N

H N

N

N

R 19 R =

H

O N H

N

N

N H

N H N

O H

N N H

N N

OC12H25 O

R

OC12H25 OC12H25

O

N

O H N R H25C12O

OC12H25 OC12H25

Figure 3.8 Photoswitchable peripheral crowding of a hexagonal rosette [23].

the periphery of the rosette, such as metal coordination, aromatic stacking, or ion pairing, in addition to peripheral crowding, and (b) to introduce more and/or more robust hydrogen bonding scaffolds. Lis porphyrin rosettes (18) [22], Yagais photoswitchable rosettes (19) [23], and Fenniris nanotubes (20) [24] nicely illustrate the former approach. Thus, CA3M3 rosettes (Figure 3.7) endowed with a pyridine moiety at the melamine component and a strapped Zn-porphyrin (to enhance coordination from only one side of the ring) at the cyanuric acid component are more stable than the CA3M3 model rosette initially reported by Whitesides. Even more complex versions of the same concept, involving one all-pyridine rosette and one all-porphyrin rosette to selfassemble into double rosettes were studied, although only NMR evidence was presented [22]. Yagai et al. have recently reported rosette 19 with an azobenzene-appended melamine and barbiturate, both possessing bulky tridodecyloxyphenyl (TDP) wedges for peripheral crowding, which is quite stable in chloroform or toluene, as established by GPC, dynamic light scattering, and NMR and UV spectroscopy [23]. It is likely UV-vis that the multiple aromatic face-to-face stacking interactions contribute to the overall stability. The geometrical EE-isomer (bearing two E-azobenzene moieties) was unusually stable to E-Z photoisomerization because of the steric crowding of the TDP wedges, whereas UV-irradiation of the monomeric melamine resulted in the ZZ isomer as expected. However, this isomer was unable to form a rosette because of the intermolecular steric hindrance between the folded TDP wedges. This is a remarkable example of photo-regulated self-assembly (Figure 3.8). Another interesting example of peripheral stabilization was provided by Fenniri [24]. Compound 20 is a DDA-AAD monomer, similar to those described by Lehn or Mascal, with a pending free amino acid that establishes further contacts via ionpairing and hydrogen bonding to stabilize the rosette and to hierarchically favor

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

Figure 3.9 Hierarchical self-assembly of rosette nanotubes from module 20. Reproduced with permission of Ref. [24].

stacking of the resulting hollow hexameric disks into nanotubular structures of ca. 4.0 nm outer diameter, that were studied by circular dichroism (CD) and transmission electron microscopy (TEM) (Figure 3.9). Another general strategy to increase the enthalpy of the intermolecular interactions is based upon the use of strong DDAA-AADD quadruple hydrogen bond motifs, such as those of 2-ureido-4-pyrimidinones (UPys), reported by Meijer and coworkers (Scheme 3.6) [28,29]. UPy dimers (21) are formed in chloroform with exceedingly high binding constants (Kdim > 107 M1), a fact that could be explained by the favorable secondary interactions between the DD-AA pairs [4,27]. Among the possible tautomeric forms that UPy can display, the pyridine-4-ol tautomer also forms strong DADA-ADAD dimers (22). Although these tautomers do not display favorable secondary interactions, their strongly acidic OH contributes to enhance the binding. UPy dimers have been used mainly to form interesting solvent-sensitive polymers [28], but also to prepare fullerene dimers [29], dendrimers [30], or polymers [31].

H O

R2 N H

O N

N R1

N H

O

R1

H N R2

21

O H

H N

N N

O

H N

R1

N H

O H N

N R2

R2 N N H O

H N

N H

22

Scheme 3.6 Tautomeric equilibria between DDAA-AADD (21) and DADA-ADAD (22) UPy dimers.

O

R1

j83

Figure 3.10 Structure of 1,3-adamantane and 2-methyl-1,3phenylene bis-UPy compounds 23 and 24 and schematic representation of the pentameric (23)5 and hexameric (24)6 assemblies [33].

84

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

It is therefore not surprising that bis-UPy scaffolds have been employed to form not only polymeric materials but also extremely stable cyclic dimers [32] and macrocycles. The ﬂexibility and shape of the linker between the two UPy subunits is a key in determining the size of the resulting dominant cyclic aggregate [33]. For example, the use of a 1,3-adamantyl spacer promotes formation of pentagonal rosettes (23)5, whereas 1,3-phenylene gives rise to hexagonal rosettes (24)6 (Figure 3.10). Compounds 23 and 24 are soluble in chloroform, dichloromethane, benzene, or toluene, but almost insoluble in methanol at 30 C. Remarkably, the viscosity of their highly concentrated (>50 mM) solutions in chloroform did not differ markedly from that of chloroform itself. Solutions of comparable concentration containing two UPys linked through ﬂexible alkyl spacers are, however, very viscous, showing that in chloroform the architecture of 23 and 24 is indeed quite different from a random coil polymer [28]. The aggregates were studied by a number of complementary techniques, such as VPO, GPC, and electrospray ionization mass spectrometry (ESI-MS). A high-intensity signal corresponding to the hexamer was observed in the ESI-MS spectrum of 24 [in THF/HCOOH (1/0.05 v/v)], showing that it forms macrocycles even in the presence of HCOOH. Signals for pentamers, tetramers, and other aggregates, were also present in the spectrum. For 23, the values obtained by VPO and GPC at high concentrations (>40 mM) fully agree with those calculated for the proposed pentameric structure. Diffusion-ordered 1 H NMR spectroscopy (DOSY) experiments were performed on solutions of 24 in CDCl3 at different concentrations. Two concentration regimes were distinguished, suggesting that, above 10 mM, hexamers of 24 aggregate into highermolecular-weight structures. The overall shapes of the cyclic aggregates are hard to estimate accurately, as the inner R groups are much too large to be accommodated inside ﬂat rosettes. Furthermore, the need to avoid eclipsed conformations around the adamantane linker in 23, as well as the presence of an inner methyl group in 24, clearly point to non-planar, tubular, or bowl-shaped conformations, respectively (Figure 3.11).

Figure 3.11 Front view representations of energy-minimized pentameric (23)5 and hexameric (24)6 assemblies.

j85

86

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly R N

H

N

H N N

H O 6

N

1N

7

N R

O

N

2

N

H N N H

H

N O

N H

H

H H N

H

N R

N H N

O

O N

R = protected ribose

R N

N

N

H N N H

N

H

N R

25

Figure 3.12 Hierarchical organization in the templated self-assembly of the G-quartets.

Macrocycles built from hydrogen bonds are not exclusive to synthetic systems. Nature offers a number of interesting examples. For instance, the G-quartet, a hydrogen-bonded macrocycle formed by cation-templated assembly of guanosine, ﬁrst identiﬁed in 1962 by Gellert et al. [34], is not only a biologically relevant DNA motif, but also a perfect example of Natures way to achieve rosettes. In the G-quartet, just two AD-DA hydrogen-bonded pairs are employed, linking the Hoogsteen and the Watson-Crick edges of guanine. The remaining carbonyl acceptor of the Watson-Crick edge is not involved in hydrogen bonding, such as in the typical G-C base pairs, but coordinates to a central metal instead, favoring cyclization and stacking of the quartets into higher-order aggregates (octamers or more complex quadruplexes) (25, Figure 3.12) [35]. A

O

O H

RO

O

RO

8

N N

H2' H1' H2'' 26-anti (A)

HN H 2N

NH N

A

NH2

N

N

RO

O

R = COC9H19 RO

8

N

H

A

S

A

H2' H1' H2''

S

26-syn (S) S S

Figure 3.13 Self-sorting of syn-anti conformations about the glycosidic bonds in Gottarellis octamer.

j87

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

R N

H N

H

N

N

H O N

N N

N R

N H

N

M+ O

N

N

H

N

N N R

N

N N

H

H H N

O

H N H

N H

O

O

M+

H N

H

O N N R

90º

N

H H N

O

N isoG

N

N

N H

N

H

N

O

N H G4-quartet R

R

R N

N R

H

H N H

N

N

N H N

H

N

N R

H N H

H N

O N

G

H H N

O

N H

H N H

H

N

O

H

R N

N

O

H N

H N H

R N

N N H

N N

N

H N N O

H

N N

N

N R

isoG5-pentamer Figure 3.14 Relative orientation of hydrogen-bonding edges (van der Waals angle) to give G4-quartets and isoG5-pentamers.

Stacking of quartets, like in DNA double helices, prevents water from competing with the hydrogen bonding network, and is thus essential in order to allow selfassembly in water. Unlike G-quartets, most of the synthetic rosettes inspired by this natural scaffold are lipophilic and were designed to form macrocycles in organic solvents. In 1995, Gottarelli et al. reported the self-assembly of 30 ,50 -didecanoyl-20 -dG (26) in chloroform [36]. Depending on conditions, 26 can extract Kþ picrate from water into chloroform to give either a discrete octamer, (26)8Kþ, or a polymer, [(26)4]n(Kþ)m. Only one stereoisomer was detected by NMR. In one G-quartet, all dG units have a syn- conformation about the glycosidic bond, while the other tetramer is in an all-anti conformation (Figure 3.13) [37]. Most remarkably, whereas metal-templated guanine self-assembly favors tetramers (G-quartets), iso-guanine tends to form cyclic pentameric arrays in the presence of a suitable metal, such as Csþ [38]. The angle between the two

67º

R

OR

27-anti

RO

O

N N

O N H

OR

N

O

27-syn

O

N

RO

H N H RO

R = COCHMe2

N H

H

Scheme 3.7 Conformational constraints in 27 promote formation of quartets without metal templating.

RO

Me2N

N

H N N

X b

NMe2

a

N

N H

Ar

H N H

N

Rib N

Ar

O

O

N Rib

N

H N H

N

Rib N N

N Ar

H N

N H N H

Gn-ribbon

N Rib

N

H N H

Rib N

N H

Ar

O

O

N

N Rib

N

H N H

Ar

H N

N Rib

N

H N H

N

Rib N

Ar

N

N

H N

Ar

O

N

O

N

H N H G4-quartet N

H

H N

H

H

N

O

O

N H

Ar

88

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

3.2 Strategies to Build up Supramolecular Macrocycles Based on Hydrogen Bonds

hydrogen-bonding edges is 90 for guanine and 67 for iso-guanine (Figure 3.14). This has also been observed even with iso-G enriched strands of DNA [39]. A general rule for self-assembly into G-quartets is the presence of a templating metal to stabilize the resulting macrocycle. However, an empty G-quartet can be obtained by careful design promoting preorganization (i.e. conformational control) and preventing aggregation into ribbon-like, undesired aggregates. Sessler et al. have reported that guanine derivative 27 forms stable cyclic quartets both in the solid state and in solution in the absence of any stabilizing cation (Scheme 3.7, path a) [40]. Ribbon structures usually arise from participation of the nitrogen at position 3 as hydrogen acceptor (Scheme 3.7, path b). This is not possible in 27 because of the steric hindrance caused by the sugar moiety, whereas the bulky aryl substituent at position 2 forces a syn conformation about the glycosidic bond, locking the structure into the correct conformation for cycle formation. An important difference between macrocyclic structures based on covalent bonding and those based on hydrogen bonds is that the latter are highly stable thermodynamically but by no means as kinetically inert as their covalent counterparts. Consequently, it has been possible to study the exchange process between molecular components of distinct cyclic aggregates. Dissociation rates of hydrogen-bonded aggregates that are not affected by mechanical constraints (steric imprisonment) are solely dependent on the number of hydrogen bonds that are broken in the dissociation process. For example, the stability constant of a dimeric assembly stabilized by two hydrogen bonds in a DA-AD array can be estimated to be around 30 M1. Assuming that the second-order complexation rate is close to diffusion control, that is, kon ¼ 1 · 108 s1 M1 for an association constant Ka ¼ kon/koff ¼ 30 M1, the ﬁrst order decomplexation rate constant could be estimated as koff ¼ 3.3 · 106 s1. This value implies that dissociation is too fast for the 1 H NMR timescale. For association constants around 106 M1 or higher the dissociation rate becomes either on the 1 H NMR timescale or even slower (koff < 100 s1). Rate constants close to 20 s1 (DG6¼ ¼ 15.7 kcal mol1) have been calculated in CDCl3 at 20 C for the dissociation of the barbituric units from assemblies of the type 143153 (Figure 3.4) [16]. This exchange process corresponds to the cleavage of six hydrogen bonds in the multimolecular assembly 143153, and, as mentioned previously, it takes place without complete disruption of the aggregate. The exchange rates of the components of cyclic and acyclic hydrogen-bonded complexes are highly sensitive to the polarity and hydrogen-bonding capabilities of the solvent. The exchange of dimelamine units (cleavage of 12 hydrogen bonds) corresponding to two different homomeric assemblies of the type 143153, detected by the formation of heteromeric assemblies, occurs within seconds in CDCl3, even at 50 C, while it takes hours at 25 C to reach the thermodynamic equilibrium in toluene. Likewise, a rate constant of 7 · 105 s1 has been reported for the dissociation of the dimelamine fragment 14 in the assembly formed with n-butyl cyanuric acid at 70 C in benzene-d6 [17]. Increasing the number of hydrogen bonds and reducing the number of components is a strategy to construct robust self-assembled macrocycles, both

j89

90

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly thermodynamically and kinetically. The data discussed previously shows how dissociation rates of components pertaining to assemblies held together by multiple hydrogen bonds decrease as the number of hydrogen bonds and the strength of the interaction increases [41]. Thus, one may expect that such hydrogen bond-induced self-assembly processes become kinetically controlled. This is a well-known phenomenon for covalent synthesis and even for self-assembly processes induced by metal-ligand coordination [42]. One leading example is provided by Reinhoudts double rosettes, where components can be exchanged in a controllable way once the rosette has been formed [16].

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

Molecular cavities have been intensively studied over the last decades by supramolecular chemists, since they can be used for molecular recognition, catalysis, or transport processes across membranes. Bowl-shaped molecules such as calixarenes (28), resorcinarenes (29), or cyclotriveratrylene (CTV, 30) derivatives constitute useful building blocks for cavities since their shapes and sizes can be controlled by covalent synthesis to a remarkable degree (Figure 3.15). The inherent ﬂexibility of calixarenes and resorcinarenes can be limited either by substitution or by bridging, respectively. Bridged resorcinarenes are known as cavitands (such as 31 or 32). However, cavitand 32 is not yet fully preorganized as it can give two main conformers, namely a cylindrical vessel or the open kite (Figure 3.16). The vase conformation can be stabilized by a seam of eight intramolecular amide hydrogen bonds forming a macrocyclic ring (compound 33) that surrounds the open R

R

R

OH

R

OH HO

OH OH OHHO

HO HO

OH OH OH

R

R

28

R

R

29 MeO

MeO

OH

HO

OH OMe

30 Figure 3.15 Representative examples of bowl-shaped cyclic oligomers of phenols.

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

O O

R X

X

X

X X

XX

O

O O

O

O

R 31

R

O

R

X

X

X

X O

O O

O

R

OR

O O

O

O

O X

X

O O

O

OO X

X R

j91

X R

R

R

R

Kite (C2v)

Vase (C4v) 32 R = C11H23; X = NO2 Figure 3.16 Cavitands as examples of preorganized macrocycles (a substantial conformational ﬂexibility remains in 32).

portal [43]. The head-to-tail cyclic array of hydrogen-bonded amides results in two cyclo-enantiomers, with clockwise and counterclockwise orientation (Figure 3.17). Because of the cyclo-enantiomerism, the C(O)CH2 protons of 33 become diastereotopic. Thus, the 1 H NMR spectrum in various non-polar solvents shows two N-H resonances shifted far downﬁeld. The chemical shifts of these signals are concentration independent, but their temperature coefﬁcients (Dd/DT) conﬁrm that one of the hydrogen bonds in 33 is less sensitive to temperature and is shielded from external solvent, while the other is somewhat more exposed. The two N-H signals coalesce in toluene-d8 at Tc 87 C. Most often, the hydrogen-bonding edges of concave surfaces have been used for self-assembly into hollow capsules. Pioneering work was related to the use of the curvature of glycoluril to build up cavities endowed with hydrogen bonding edges at the concave and convex inner sides, and outer ends that are complementary and promote formation of pseudo-spherical cavities such as 34 (tennis ball) or 35 (soft ball) (Figure 3.18) when in the presence of suitable guests [44]. The subject has been extensively studied by Rebek and coworkers, who have written a substantial number of reviews covering different aspects of the topic [45]. Because of this, we do not intend to describe here in detail the different systems that have been developed, but focus instead on some of the dynamic aspects. Possible applications of these systems include molecular containers, enantioselective receptors, catalysts, drug-delivery devices, separation tools or simply unusual

Figure 3.17 Cyclic array of hydrogen-bonded amides resulting in cyclo-enantiomerism.

92

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

Figure 3.18 Glycoluril receptors 34 and 35 (R1 ¼ R2 ¼ CO2-ipentyl, R3 ¼ H) self-assemble into tennis-ball and soft-ball dimers, respectively.

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

j93

94

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

Scheme 3.8 A Diels-Alder reaction is accelerated in a tennis-ball like hydrogen bonded capsule (35)2 (R1 ¼ R2 ¼ 4-n-heptylphenyl, R3 ¼ OH). True catalysis is prevented by product inhibition.

closed spaces where the encapsulated guests can display emerging properties or unexpected behavior. In contrast to capsules based on covalent bonds (such as Crams carcerands) [46], whose portals are usually too small to allow fast exchange of guests, hydrogen bonded self-assembly is beneﬁcial for encapsulation, as it is reversible. For large capsules or multicomponent assemblies (see below), the guests can access the cavities by partial disruption of the hydrogen bonding array (opening of hosts ﬂaps) or by separation of one of the components. Smaller dimeric capsules in general, however, require full dissociation of the two components. The process is relatively slow, on the order of seconds, although in extreme cases it can take from milliseconds to hours. Despite some early successful reports, there are concerns about the practical use of self-assembled vessels for catalysis by encapsulation, especially for reactions that are inherently faster than the opening-closing events. For instance, a Diels-Alder reaction can be accelerated in the presence of the soft-ball capsule 35 (Scheme 3.8), although product inhibition forces the use of large amounts of capsule in the reaction mixture [47].

Scheme 3.9 Hydrogen-bonded dimeric capsule (35)2 (R1 ¼ R2 ¼ 4-n-heptylphenyl, R3 ¼ OH) catalyzes a Diels-Alder reaction.

O

H

N N

O

O

O

H

H N

O

H

O

O

O

O

H O

O H

O

R

R

36

R

R

O

O O H

N H

O

O

H N

N

H

H

O

N

O

O O

H

R = C11H 23 H

O

O

H

N

N

N

N

N

O

O

O

N

O

Figure 3.19 Cylindrical capsules from imides and ureas at the bridges of cavitands.

O

O

H

H N

O

(36)22 (36)

O

O

37

R

R

1.8 nm

R

R

O

O

N

N

N

O

O

O

N

O

N

N

H

O

O

H

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

j95

96

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

Scheme 3.10 Closed spaces inside a capsule impose restrictions in the orientation of reaction partners in a 1,3-dipolar cycloaddition.

For the more favorable case depicted in Scheme 3.9, where the Diels-Alder adduct is bound less tightly than the reactants, true catalysis and some turnover was observed [48]. In both cases, however, productive capsules (i.e. ﬁlled with diene and dienophile components) have to compete additionally with non-productive ones (ﬁlled with two dienes or two dienophiles). The problem of correctly ﬁlling the capsules with only productive partners has been addressed by cylindrical capsules, such as (36)2 [49] or (37)2 [50], which can simultaneously hold one large and one small guest, but not two large (not enough room) or two small (too much void space left) ones. For instance, (36)2 accommodates pair-wise carboxylic acid dimers (Figure 3.19) [49]. In addition, cylindrical shapes offer the possibility to favor the alignment of reactants in a speciﬁc way, not favored outside of a conﬁned space. Thus, the 1,3-dipolar cycloaddition between an alkyne and an azide (Scheme 3.10) yields mainly triazole adduct 38 in the presence of (37)2, whereas a mixture of regioisomers 38 and 39 are obtained in its absence [51]. Chirality can be introduced in capsules by substitution. Examples of chiral recognition inside the cavities are scarce as introduction of asymmetry in the periphery does not translate inside the cavity. Thus, introduction of different substituents at both side glycolurils in tennis balls and soft balls, such as in 35 (R1 ¼ pC6H4OC6H15, R2 ¼ CO2-i-pentyl, R3 ¼ H), removes one of the two symmetry planes

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

Scheme 3.11 Chirality in a heterodimeric ureidocalix[4]arene capsule, induced by the orientation of the cyclic array of hydrogen bonds.

of the monomer and leaves the dimer with a C2 axis of symmetry [52]. Similarly, well known urea-urea hydrogen-bonding patterns have been independently studied by Rebek [45c,53] and Böhmer [54], using calix[4]arenes containing ureas at their wider rims. Heterodimers arising from a mixture of calix[4]arene ureas 40 and 41 are thus chiral (Scheme 3.11). Spacer 2

Spacer 1 O

O

H N

N

H N

N

N N

R

N N O

O

O

O

O

N

N H

N

N H

R

O

42 R = 4-n-heptylphenyl

x2

Achiral Monomer

Enantiomeric Dimeric Capsules

Scheme 3.12 Chirality generated by dimerization of an achiral monomer with two different spacers.

j97

98

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly None of these dissymmetric capsules display enantioselectivity for chiral substrates. On the contrary, more distorted scaffolds, such as those resulting from use of different spacers in a soft ball (compound 42 in Scheme 3.12) yield enantiomeric capsules that can discriminate chiral guests, with a modest diastereomeric excess (up to 35% for the most favorable cases) [55]. When an enantiopure chiral guest is added to a racemic softball mixture, the initial 1:1 kinetic distribution of diastereoisomeric ﬁlled capsules slowly turns into a thermodynamic distribution favoring one diastereoisomer, on the timescale of capsule dissociation (t1/2 ¼ 19 h). If an excess of the less favored enantiomer is added rapidly, a distribution favoring this unstable complex shows up quickly, and then the thermodynamic mixture is slowly reached again upon standing. This memory effect (one capsule retains its chirality even when its shaping guest template is removed) indicates also that complete dissociation of the capsule is unlikely for the exchange, and a ﬂap-opening mechanism operates instead [56]. Exchange in cylindrical capsules, such as (37)2, has also been studied by Rebek and coworkers. The lifetime of the capsule is ca. 0.5 s, and up to three guests can be accommodated inside, depending on their sizes and the solvent employed. Even for larger single guests, such as 4,40 -dimethylbiphenyl, exchange may occur without dissociation of the capsule and without formation of a completely vacant capsule. Most remarkably, n-alkanes that are much longer than the capsule can be accommodated, ntetradecane being the longest guest allowing encapsulation. To achieve this, the alkane must coil into a chiral helical conformation. Detailed NMR studies show that the helix rapidly racemizes inside the capsule without fully unwinding, which suggests that racemization proceeds through propagation of short-range uncoiling [57]. Exchange processes in ureidocalix[4]arenes (i.e. 40 or 41) have also been studied in detail. In each monomer of these aggregates (which require a guest the size of benzene to properly form), ureas are oriented in opposite directions (P or M), and the direction of rotation of the cyclic array is slow at the NMR timescale. The capsule is therefore a meso form. Exchange of benzene guests in these calixarene dimers occurs by simultaneous dissociation coupled with rotation of the ureas (racemization) [58]. For cationic guests, like tetraethylammonium, the bulkier substrate forces the capsule to expand to some extent relative to benzene. This weakens some hydrogen bonds (which is compensated by cation-p host-guest interactions), and favors rotation of the ureas to racemize without fully disrupting the dimeric assembly [59]. Wider ureidocalixarenes, such as calix[5]- or calix[6]arenes, cannot dimerize since there is no space between the ureas to allow interpenetration with the set of ureas provided by the second monomer, but calix[6]arene 43, with just three unsubstituted ureas at alternate rings, forms quite stable dimers [60]. Interestingly, addition of increasing amounts DMSO to a 1:1 mixture of a triureidocalix[6]arene and tetraureidocalix[4]arene capsules in chloroform causes disruption of the calix[4]arene capsule (containing 8 ureas) prior to the calix[6]arene ones (endowed with a seam of just 6 ureas), a fact that can be explained by the better arrangement of the hydrogen bonding arrays in the latter (Scheme 3.13). The ﬂexibility and mutability of hydrogen bonds, as compared to covalent bonds, make self-assembled capsules unlikely to display good selectivities for closely related

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

Scheme 3.13 Dimeric capsule formation in a triureidocalix[6]arene.

guests. A remarkable exception has recently been found in our laboratory for the encapsulation of fullerenes inside capsule (44)2 built up from two CTV scaffolds endowed with three UPy hydrogen-bonding arrays [61]. This large capsule was speciﬁcally designed to encapsulate fullerenes, and indeed both C60 and C70 were readily complexed by (44)2, with binding constants in the order of 103 and 104 M1, respectively. The selectivity for C70 allows for an expeditious method, not relying on chromatography, to efﬁciently extract this valuable material from carbon soot or from fullerite, which was attributed to the slightly different shape of each respective fullerene. Whereas C60 is roughly spherical, like a soccer ball, C70 is thinner and slightly elongated, resembling a rugby ball. The hydrogen-bond patterns of the UPy dimeric quadruple bond arrays are closer to the ideal planarity in the C70 complex (Figure 3.20). Examples of enclosing the space with aggregates composed of more than two molecular components and simply stabilized by hydrogen bonds are scarce in the literature [62]. Cyclic sulfonamides (such as 45) feature a self-complementary hydrogen bonding pattern of glycolurils (Scheme 3.14). Under the appropriate conditions, sulfonamides and glycolurils like to form heteromeric hydrogen-bonding aggregates, as we have previously seen demonstrated by the heterodimeric ureidocalix[4]arene dimeric capsules. Curved monomer 46 [62], containing both groups, is thus programmed to self-assemble in a head-to-tail manner, establishing hydrogen bonds between the best two partners. The ﬁnal result is the assembly of a rare tetramolecular capsule (46)4, induced by encapsulation of guests of appropriate size and functionality, such as adamantane (Figure 3.21). Adamantane (47), however, only offers weak van der Waals interactions to the inner concave surface of the host assembly. This interaction, together with the enthalpic gains provided by the formation of 16 hydrogen bonds, is enough to overcome the entropy penalty for bringing together four monomers and one guest.

j99

Figure 3.20 Dihedral angles closer to planarity between UPy binding subunits account for selective extraction of C70 with cyclotriveratrylene-based capsule (44)2.

100

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds R N N O

O

H

H R N N

N N R H

H

H O N S O N H 45

O

j101

O N

N R

H O N S O N H

H R N H N O

O N

N R

H

N O S N O H

Scheme 3.14 Heterodimers formed between sulfonamides (R ¼ 4-n-heptylphenyl) and glycolurils.

Adamantane-2,6-dione (48) further stabilizes the tetrameric assembly by the presence of hydrogen-bonding acceptor groups (C¼O) in the guest molecules through the formation of bifurcated hydrogen bonds with the glycoluril NH groups of the capsule (Figure 3.22). Dimeric self-assemblies based on resorcinarene cavitands have already been mentioned. However, resorcin[4]arene itself (49a) was the subject of the ﬁrst report of a selfassembly process, directed simply by hydrogen bonds, capable of yielding a giant hexameric capsule. Thus, in 1997 Atwood and MacGillivray showed that 49a crystallized into a chiral spherical molecular assembly [62b]. Six monomers (49a) assembled together, with eight structural water molecules occupying the corners of the assembly, to form a supramolecular hexameric structure that conforms to a snub cube (Figure 3.23). The capsule is held together by a complex pattern of 60 OH H hydrogen bonds, 24 of which are intramolecular and help to stabilize the crown-like conformation of each resorcinarene component. Consequently, the six monomers and the eight water molecules are held together by 36 intermolecular hydrogen bonds (HB/(N 1) ¼ 2.8). The enclosed space in the cavity of the hexamer (49a)6 is about 1403 Å3 [63]. As expected, the solvent molecules present inside the cavity of (49a)6, in the solid state, were too disordered to be located. Authors showed preliminary supporting evidence that the hexameric structure was also maintained in solution.

Figure 3.21 A glycoluril with a sulfonamide self-assembles into a tetramer.

102

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly

Figure 3.22 Encapsulation of adamantane (left) and adamantane-2,6-dione (right).

Figure 3.23 The resorcinarene octahedral self-assembled hexamer (left) and the volume of the enclosed cavity (right).

3.3 Strategies to Build up Supramolecular Cavities and Capsules Based on Hydrogen Bonds

j103

Crystal structures for several of the corresponding pyrogallol[4]arenes 50 (Figure 3.23) have been solved. These also form hexameric capsules, which now contain an array of 72 hydrogen bonds (48 are intramolecular and 24 intermolecular, HB/ (N 1) ¼ 4) and have no structural water molecules [64]. The increase in the ITm HB/(N 1) value agrees very well with the fact that nanocapsules based on pyrogallol 50a are more stable in polar media than those derived from resorcinarene 49a. A decade ago Aoyama et al. described the formation of a simple 1:1 open complex of resorcinarene 50b with dicarboxylic acids in CDCl3 and, subsequently, with ribose, terpenes and even steroids [65]. Convincing evidence from laboratories worldwide indicate that resorcinarenes 49 also assemble in water-saturated organic solvents as a hexamer [66]. These hexameric assemblies are observed in the encapsulation of a wide variety of guests in solution. However, the pyrogallol[4]arene hexamer does not require water to be stabilized and assembles in dry halogenated solvents as well. Addition of certain neutral [67] and positively charged [68] guests leads to an upﬁeld shifted signals for the encapsulated guests protons in the 1H NMR spectra. The protons of the host experience do not exhibit signiﬁcant changes in chemical shift. The exchange of the free and encapsulated guest is slow on the NMR timescale, as indicated by the appearance of two sets of resonances. Most guests occupy only about 55% of the available volume [69]. Interestingly, many guests within hexameric capsules show packing coefﬁcients (PC) similar to those in solution, viz. ca. 0.5 [PC ¼ volume of the guest(s)/volume of the cavity (1400 Å3)] [69]. Recently, Schalley and coworkers have presented evidence for the hexameric capsular structure, even in the gas phase after ionization [70]. The molecular structure of ruthenium complex 512þ is congruent with the size, shape, and symmetry of the cavity of the hexameric host (49a)6. Molecular modeling studies show an almost ideal ﬁtting of 512þ inside the cavity. The periphery of each pyridine ring on the guest cation is directed toward each end of the six resorcinarene monomers. Although the guest is only pseudo octahedral and deviates somewhat from a perfect octahedral geometry, it nevertheless matches well enough the structure of the hexameric capsular host. The

Figure 3.24 Ruthenium complex 51 is encapsulated in the hexameric self-assembly of resorcinarene 49.

104

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly OH O

52 Figure 3.25 Pyrene butyric acid, a ﬂuorescent guest to study interactions inside a capsule.

ESI mass spectra after the addition of 49 (or 50) to 51(PF6)2 show the hexameric clusters 51@(49)6 and 51@(50)6 as the dominating signals (Figure 3.24). These ions were irradiated in an infrared multiphoton dissociation experiment (IRMPD) with IR radiation from a CO2 laser at different time intervals. The observed fragmentation behavior was exactly that expected for the hexameric capsule. To further explore the environment inside such hexameric hosts, Atwood and coworkers performed structural characterization and spectroﬂuorometric studies of the capsule-bound ﬂuorescent probe molecule pyrene butyric acid (PBA) 52 [71]. Single-crystal X-ray diffraction studies on 52@(50b)6 show not only that encapsulated guests interact with the host walls in the solid state, but also that the p surfaces of the guest molecules are well separated from one another within the capsule. The spectroscopic studies in solution corroborated this ﬁnding and revealed an average of 1.5 molecules of 52 per capsule. The assembly remains intact over four weeks in the solution phase, suggesting that the carboxylic acid groups and the polyaromatic nature of 52 do not destabilize the overall supramolecular assembly. Resorcinarene platforms have also been developed in the generation of multicomponent capsules stabilized by hydrogen bonds. In this sense, Kobayashi et al. have constructed a methylene-bridged resorcinarene scaffold functionalized with four carboxylic acids (53) [72]. 2-Aminopyrimidine is used as a wedged-shaped hydrogen-bonding bridge to form two hydrogen bonds with each of two carboxylic acids on neighboring molecules (Figure 3.26).

Figure 3.26 Two cavitands containing four carboxylic acids dimerize assisted by 2-aminopyrimidine hydrogen-bonded connectors.

3.4 Summary and Outlook

j105

Figure 3.27 An expanded, six-component capsule is formed by addition of glycoluril 54 to cavitand 37, which would otherwise form a dimeric (37)2 capsule.

Rebek and Ajami have recently reported the reversible expansion of the dimeric capsule (37)2 with four glycouril spacers (54), increasing the cavitys volume by 200 Å3 and its length by 7 Å [73]. The expanded capsule self-assembles from six molecular components and is stabilized by 32 intermolecular hydrogen bonds. The expanded capsule is capable of encapsulating normal alkanes up to C21H44 and other guests of similar length, such as the biphenyl-alkyne derivative 55 shown in Figure 3.27.

3.4 Summary and Outlook

The non-covalent synthesis of discrete macrocyclic and macrooligocyclic structures using multiple hydrogen-bonding interactions has now become a reality. The supramolecular approach outlined here represents an attractive and effective alternative to traditional stepwise covalent synthesis. Use of multiple hydrogen bonds to organize molecular components into supramolecular nanostructures has

106

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly revolutionized the way in which chemists approach the construction of complex cyclic structures. The examples discussed in this chapter take place under thermodynamic control. This means that the products must be evaluated in terms of thermodynamic minima in equilibrating mixtures. As a consequence, they reﬂect a balance between enthalpy and entropy. We have presented and discussed several principles that can be used to inﬂuence the association of the components toward cyclic aggregates and to predict their relative stabilities. The stability of a cyclic multimolecular aggregate stabilized by intermolecular interactions is not simply a consequence of DG , since this thermodynamic variable does not take into account the effect of the concentration (entropy) on the extent of the aggregation. The concepts derived from the study of simple model systems of macrocyclic structures stabilized by hydrogen bonds could be used to rationalize the association of more complex biological aggregates. In addition, the construction of cyclic aggregates stabilized by hydrogen bonds paved the way for the study of reversible molecular encapsulation and the nature of intimate molecular relationships isolated from the bulk solvent, as well as the effects that molecular conﬁnement in nanoscopic spaces has chemical reactivity.

3.5 Experimental: Selected Procedures 3.5.1 Solid State Formation of the Hexameric Capsule Derived from Pyrogallol[4]arene (50c) [74]

Compound 50c was obtained in ca. 25% yield as a precipitate from the acid-catalyzed condensation of pyrogallol and isovaleraldehyde. No evidence of any hexamer was found in the solid material. To convert this material into the hexamer (50c)6, the original precipitate can be dissolved in Et2O, acetone, or methanol, with a few drops of nitrobenzene or o-dinitrobenzene, followed by crystallization upon slow evaporation. The hexamer may also be obtained by thermal treatment of the initial precipitate or the initial ﬁltrate. The product in the initial ﬁltrate may be converted into hexamer by extraction in Et2O, followed by evaporation to dryness with subsequent dissolution in methanol. The methanol solution is then heated to 120–150 C for at least 12 h. Methanol may be removed under vacuum to yield a red-brown solid. Colorless hexameric spherical capsules are obtained from this solid utilizing the crystallization procedure described for the initial precipitate. 3.5.2 Crystals of the Host-Guest Arrangement of 52@(50b)6

A hot, saturated HPLC grade acetonitrile solution containing pyrene butyric acid (52, 2 equivalents) was added to a hot, saturated HPLC grade acetonitrile solution

Abbreviations

j107

containing C-hexylpyrogallol[4]arene (50b) to give a pale yellow solution. The solution was sonicated for 10–15 min, re-heated, and allowed to cool slowly. The sample vial was capped and allowed to stand overnight, after which time the solution had darkened to a brown colour. Slow evaporation and concomitant concentration of this solution over a two to three week period afforded the growth of pale single crystals suitable for X-ray diffraction studies. The crystals were ﬁltered off and used in the spectroﬂuorometric studies as required.

Acknowledgments

Financial support from MEC (CTQ2005-08948-C02-01/BQU, CTQ2005-08989-0102/BQU, and CSD2006-0003), ICIQ Foundation, and Generalitat de Catalunya (2005SGR00108) is gratefully acknowledged.

Abbreviations

EM NMR CA M DMSO A D VPO GPC G C TDP UV CD TEM Upy ESI MS THF DOSY CTV PC IRMPD PBA HPLC

effective molarity nuclear magnetic resonance cyanuric acid melamine dimethylsulfoxide acceptor donor vapor pressure osmometry gel permeation chromatography guanine cytosine tridodecyloxyphenyl ultraviolet circular dichroism transmission electron microscopy 2-ureido-4-pyrimidinone electrospray ionization mass spectrometry tetrahydrofuran diffusion ordered spectroscopy cyclotriveratrylene packing coefﬁcients infrared multiphoton dissociation pyrene butyric acid high performance liquid chromatography

108

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly References 1 (a) Ercolani, G. (2003) J. Am. Chem. Soc., 125, 16097–16103; (b) Ercolani, G. (2003) J. Phys. Chem. B, 107, 5052–5057. 2 Hunter, C.A. (2004) Angew. Chem., 116, 5424–5439; Angew. Chem. Int. Ed., 43, 5310–5324. 3 (a) Jeong, K.S., Tjivikua, T. and Rebek, J., Jr (1990) J. Am. Chem. Soc., 112, 3215–3217; (b) Sartorius, J. and Schneider, H.-J. (1996) Chem. Eur. J., 2, 1446–1452. 4 Jorgensen, W.L. and Pranata, J. (1990) J. Am. Chem. Soc., 112, 2008–2010. 5 Ercolani, G., (ed. E. Alessio) (2006) Noncovalent Multiporphyrin Assemblies, Springer, Berlin. 6 Zimmerman, S.C. and Duerr, B.F. (1992) J. Org. Chem., 57, 2215–2217. 7 Ducharme, Y. and Wuest, J.D. (1988) J. Org. Chem., 53, 5787–5789. 8 Seto, C.T. and Whitesides, G. (1993) J. Am. Chem. Soc., 115, 905–916. 9 Whitesides, G.M., Simanek, E.E., Mathias, J.P., Seto, C.T., Chin, D., Mammen, M. and Gordon, D.M. (1995) Acc. Chem. Res., 28, 37–44. 10 Mammen, M., Simanek, E.E. and Whitesides, G.M. (1996) J. Am. Chem. Soc., 118, 12614–12623. 11 Experimental entropy measurements include contributions from desolvation and changes in internal structure: (a) Kyogoku, Y., Lord, R.C. and Rich, A. (1969) Biochim. Biophys. Acta, 179, 10–17; (b) Searle, M.S. and Williams, D. H. (1992) J. Am. Chem. Soc., 114, 10690– 10697. 12 A theoretical estimate of the entropy of association for the 6-particle aggregate reported by Whitesides et al. is in the order of 50 kcal mol1: Mammen, M., Shakhnovich, E.I., Deutch, J.M. and Whitesides, G.M. (1998) J. Org. Chem., 63, 3821–3830. 13 Cram, D.J. (1986) Angew. Chem., 98, 1041–1060; Angew. Chem. Int. Ed. Engl., 25, 1039–1057.

14 Mathias, J.P., Simanek, E.E., Zerkowski, J.A., Seto, C.T. and Whitesides, G.M. (1994) J. Am. Chem. Soc., 116, 4316– 4325. 15 (a) Chi, X., Guerin, A.J., Haycock, R.A., Hunter, C.A. and Sarson, L.D. (1995) J. Chem. Soc. Chem. Commun., 2563–2565; (b) For a complete theoretical treatment of self-assembly macrocyclizations under thermodynamic control see also: Ercolani, G. (1998) J. Phys. Chem. B, 102, 5699–5703. 16 For a review, see: Prins, L.J., Reinhoudt, D.N. and Timmerman, P. (2001) Angew. Chem., 113, 2446–2492; Angew. Chem. Int. Ed., 40, 2382–2426. 17 Prins, L.J., De Jong, F., Timmerman, P. and Reinhoudt, D.N. (2000) Nature, 408, 181–184. 18 (a) Prins, L.J., Huskens, J., de Jong, F., Timmerman, P. and Reinhoudt, D.N. (1999) Nature, 398, 498–502; (b) Prins, L. J., Hulst, R., Timmerman, P. and Reinhoudt, D.N. (2002) Chem. Eur. J., 8, 2288–2301; (c) Prins, L.J., Verhage, J.J., de Jong, F., Timmerman, P. and Reinhoudt, D.N. (2002) Chem. Eur. J., 8, 2302–2313. 19 (a) Yashima, E., Maeda, K. and Okamoto, Y. (1999) Nature, 399, 449–451; (b) Furusho, Y., Kimura, T., Mizuno, Y. and Aida, T. (1997) J. Am. Chem. Soc., 119, 5267–5268. 20 (a) Marsh, A., Nolen, E.G., Gardinier, K. M. and Lehn, J.-M. (1994) Tetrahedron Lett., 35, 397–400; (b) Marsh, A., Silvestri, M. and Lehn, J.-M. (1996) Chem. Commun., 1527–1528. 21 (a) Mascal, M., Hext, N.M., Warmuth, R., Moore, M.H. and Turkenburg, J.P. (1996) Angew. Chem., 108, 2348–2350; Angew. Chem. Int. Ed. Engl., 35, 2204–2206; (b) Mascal, M., Hext, N.M., Warmuth, R., Arnall-Culliford, J.R., Moore, M.H. and Turkenburg, J.P. (1999) J. Org. Chem., 64, 8479–8484.

References

22 Shao, X.-B., Jiang, X.-K., Zhu, S.-Z. and Li, Z.-T. (2004) Tetrahedron, 60, 9155–9162. 23 Yagai, S., Nakajima, T., Karatsu, T., Saitow, K. and Kitamura, A. (2004) J. Am. Chem. Soc., 126, 11500–11508. 24 Fenniri, H., Mathivanan, P., Vidale, K.L., Sherman, D.M., Hallenga, K., Wood, K. V. and Stowell, J.G. (2001) J. Am. Chem. Soc., 123, 3854–3855. 25 (a) Sijbesma, R.P., Beijer, F.H., Brunsveld, L., Folmer, B.J.B., Hirschberg, J.H.K.K., Lange, R.F.M., Lowe, J.K.L. and Meijer, E.W. (1997) Science, 278, 1601–1604; (b) Beijer, F.H., Sijbesma, R.P., Kooijman, H., Spek, A.L. and Meijer, E.W. (1998) J. Am. Chem. Soc., 120, 6761–6769; (c) Hirschberg, J. H.K.K., Brunsveld, L., Ramzi, A., Vekemans, J.A.J.M., Sijbesma, R.P. and Meijer, E.W. (2000) Nature, 407, 167–170; (d) Brunsveld, L., Vekemans, J. A.J.M., Hirschberg, J.H.K.K., Sijbesma, R.P. and Meijer, E.W. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 4977–4982; (e) Ligthart, G.B.W.L., Ohkawa, H., Sijbesma, R.P. and Meijer, E.W. (2005) J. Am. Chem. Soc., 127, 810–811. 26 For similar motifs based on other heterocyclic systems, see for instance: (a) Zimmerman, S.C., Zeng, F., Reichert, D.E.C. and Kolotuchin, S.V. (1996) Science, 271, 1095–1098; (b) Kolotuchin, S.V. and Zimmerman, S.C. (1998) J. Am. Chem. Soc., 120, 9092–9093; (c) Corbin, P.S., Lawless, L. J., Li, Z., Ma, Y., Witmer, M.J. and Zimmerman, S.C. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 5099–5104; (d) Lüning, U., Kühl, C. and Uphoff, A. (2002) Eur. J. Org. Chem., 4063– 4070; (e) Brammer, S., Lüning, U. and Kühl, C. (2002) Eur. J. Org. Chem., 4054–4062. 27 Pranata, J., Wierschke, S.G. and Jorgensen, W.L. (1991) J. Am. Chem. Soc., 113, 2810–2819. 28 Folmer, B.J.B., Sijbesma, R.P., Versteegen, R.M., van der Rijt, J.A.J. and

29

30

31

32

33

34

35

36

37

38

j109

Meijer, E.W. (2000) Adv. Mater., 12, 874–878. (a) Rispens, M.T., Sánchez, L., Knol, J. and Hummelen, J.C. (2001) Chem. Commun., 161–162; (b) González, J.J., González, S., Priego, E.M., Luo, C., Guldi, D.M., de Mendoza, J. and Martín, N. (2001) Chem. Commun., 163–164. Hahn, U., González, J.J., Huerta, E., Segura, M., Eckert, J.-F., Cardinali, F., de Mendoza, J. and Nierengarten, J.-F. (2005) Chem. Eur. J., 11, 6666–6672. See, for instance: Rispens, M.T., Sánchez, L., Beckers, E.H.A., Van Hal, P. A., Schenning, A.P.H.J., El-ghayouri, A., Peeters, E., Meijer, E.W., Janssen, R.A.J. and Hummelen, J.C. (2003) Synth. Metals, 135-136, 801–803, and references therein. Folmer, B.J.B., Sijbesma, R.P., Kooijman, H., Spek, A.L. and Meijer, E. W. (1999) J. Am. Chem. Soc., 121, 9001– 9007. Keizer, H.M., González, J.J., Segura, M., Prados, P., Sijbesma, R.P., Meijer, E.W. and de Mendoza, J. (2005) Chem. Eur. J., 11, 4602–4608. Gellert, M., Lipsett, M.N. and Davies, D. R. (1962) Proc. Natl. Acad. Sci. USA, 48, 2013–2018. For an excellent recent review on G-quartets, see: Davis, J.F. (2004) Angew. Chem., 116, 684–716; Angew. Chem. Int. Ed., 43, 668–698. Gottarelli, G., Masiero, S. and Spada, G. P. (1995) J. Chem. Soc. Chem. Commun., 2555–2557. Marlow, A.L., Mezzina, E., Spada, G.P., Masiero, S., Davis, J.T. and Gottarelli, G. (1999) J. Org. Chem., 64, 5116– 5123. (a) Cai, M., Marlow, A.L., Fettinger, J.C., Fabris, D., Haverlock, T.J., Moyer, B.A. and Davis, J.T. (2000) Angew. Chem., 112, 1339–1341; Angew. Chem. Int. Ed., 39, 1283–1285; (b) Shi, X., Fettinger, J.C., Cai, M. and Davis, J.T. (2000) Angew. Chem., 112, 3254–3257; Angew. Chem. Int. Ed., 39, 3124–3127.

110

j 3 Supramolecular Macrocycle Synthesis by H-bonding Assembly 39 Chaput, J.C. and Switzer, C. (1999) Proc. Natl. Acad. Sci. USA, 96, 10614–10619. 40 Sessler, J.L., Sathiosatham, M., Doerr, K., Lynch, V. and Abboud, K.A. (2000) Angew. Chem., 112, 1356–1359; Angew. Chem. Int. Ed., 39, 1300–1303. 41 A binding strength of Ka ¼ 2 107 M1 is reported for a triple hydrogen bonding array AAA-DDD: Djurdjevic, S., Leigh, D.A., McNab, H., Parsons, S., Teobaldi, G. and Zerbetto, F. (2007) J. Am. Chem. Soc., 129, 476–477. 42 Hasenknopf, B., Lehn, J.-M., Boumediene, N., Leize, E. and Van Dorsselaer, A. (1998) Angew. Chem., 110, 3458–3460; Angew. Chem. Int. Ed., 37, 3265–3268. 43 (a) Rudkevich, D.M., Hilmersson, G. and Rebek, J., Jr (1997) J. Am. Chem. Soc., 119, 9911–9912; (b) Rudkevich, D. M., Hilmersson, G. and Rebek, J., Jr (1998) J. Am. Chem. Soc., 120, 12216–12225. 44 (a) Wyler, R., de Mendoza, J. and Rebek, J., Jr (1993) Angew. Chem., 105, 1820–1821; Angew. Chem. Int. Ed. Engl., 32, 1699–1701; (b) Meissner, R.S., Rebek, J., Jr and de Mendoza, J. (1995) Science, 270, 1485–1488. 45 See for instance: (a) Morgan Conn, M. and Rebek, J., Jr (1997) Chem. Rev., 97, 1647–1668; (b) Rudkevich, D.M. and Rebek, J., Jr (1999) Eur. J. Org. Chem., 1991–2005; (c) Rebek, J., Jr (2000) Chem. Commun., 637–643; (d) Hof, F., Craig, S. L., Nuckolls, C. and Rebek, J., Jr (2002) Angew. Chem., 114, 1556–1578; Angew. Chem. Int. Ed., 41, 1488–1508; (e) Palmer, L.C. and Rebek, J., Jr (2004) Org. Biomol. Chem., 2, 3051–3059; (f) Rebek, J., Jr (2005) Angew. Chem., 117, 2104–2115; Angew. Chem,. Int. Ed., 44, 2068–2078; (g) Biros, S.M. and Rebek, J., Jr (2007) Chem. Soc. Rev., 36, 93–104. 46 (a) Morán, J.R., Karbach, S. and Cram, D.J. (1982) J. Am. Chem. Soc., 104, 5826–5828; (b) Sherman, J.C., Knobler, C.C. and Cram, D.J. (1991) J. Am. Chem. Soc., 113, 2194–2204.

47 Kang, J. and Rebek, J., Jr (1997) Nature, 385, 50–52. 48 Kang, J., Santamaría, J., Hilmersson, G. and Rebek, J., Jr (1998) J. Am. Chem. Soc., 120, 7389–7390. 49 Ebbing, M.H.K., Villa, M.J., Valpuesta, J. M., Prados, P. and de Mendoza, J. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 4962–4966. 50 Heinz, T., Rudkevich, D.M. and Rebek, J., Jr (1998) Nature, 394, 764–766. 51 Chen, J. and Rebek, J., Jr (2002) Org. Lett., 4, 327–329. 52 Tokunaga, Y. and Rebek, J., Jr (1998) J. Am. Chem. Soc., 120, 66–69. 53 (a) Shimidzu, K. and Rebek, J., Jr (1995) Proc. Natl. Acad. Sci. U.S.A., 92, 12403–12407; (b) Hamann, B., Shimizu, K.D. and Rebek, J., Jr (1996) Angew. Chem., 108, 1425–1427; Angew. Chem. Int. Ed. Engl., 35, 1326–1329. 54 (a) Mogck, O., Böhmer, V. and Vogt, W. (1996) Tetrahedron, 52, 8489–8496; (b) Mogck, O., Paulus, E.F., Böhmer, V., Thondorf, I. and Vogt, W.J. (1996) Chem. Commun., 2533–2534. 55 Rivera, J.M., Martín, T. and Rebek, J., Jr (1998) J. Am. Chem. Soc., 120, 819–820; (b) Rivera, J.M., Martín, T. and Rebek, J., Jr (1998) Science, 279, 1021–1023. 56 Rivera, J.M., Craig, S.L., Martín, T. and Rebek, J., Jr (2000) Angew. Chem., 112, 2214–2216; Angew. Chem. Int. Ed., 39, 2130–2132. 57 Scarso, A., Trembleau, L. and Rebek, J., Jr (2004) J. Am. Chem. Soc., 126, 13512– 13518. 58 Mogck, O., Pons, M., Böhmer, V. and Vogt, W. (1997) J. Am. Chem. Soc., 119, 5706–5712. 59 Vysotsky, M.O., Pop, A., Broda, F., Thondorf, I. and Böhmer, V. (2001) Chem. Eur. J., 7, 4403–4410. 60 González, J.J., Ferdani, R., Albertini, E., Blasco, J.M., Arduini, A., Pochini, A., Prados, P. and de Mendoza, J. (2000) Chem. Eur. J., 6, 73–80. 61 Huerta, E., Metselaar, G.A., Fragoso, A., Santos, E., Bo, C. and de Mendoza, J.

References

62

63 64

65

66

(2007) Angew. Chem., 119, 206–209; Angew. Chem. Int. Ed., 46, 202–205. (a) Martín, T., Obst, U. and Rebek, J., Jr (1998) Science, 281, 1842–1845; (b) MacGillivray, L.R. and Atwood, J.L. (1997) Nature, 389, 469–472. DeepView/Swiss-PdbViewer, v. 3.7.: http://www.expasy.org./spdv. (a) Gerkensmeier, T., Iwanek, W., Agena, C., Fröhlich, R., Kotila, S., Näther, C. and Mattay, J. (1999) Eur. J. Org. Chem., 1999, 2257–2262; (b) Cave, G.W.V., Antesberger, J., Barbour, L.J., McKinlay, R.M. and Atwood, J.L. (2004) Angew. Chem., 116, 5375–5378; Angew. Chem. Int. Ed., 43, 5263–5266. Aoyama, Y., Tanaka, Y., Toi, H. and Ogoshi, H. (1988) J. Am. Chem. Soc., 110, 634–635. (a) Wilson, C.F., Eastman, M.P. and Hartzell, C.J. (1997) J. Phys. Chem. B, 101, 9309–9313; (b) Avram, L. and Cohen, Y. (2002) J. Am. Chem. Soc., 124, 15148–15149; (c) Shivanyuk, A. and Rebek, J., Jr (2003) J. Am. Chem. Soc., 125, 3432–3433; (d) Evan-Salem, T., Baruch, I., Avram, L., Cohen, Y., Palmer, L.C. and Rebek, J., Jr (2006)

67 68

69 70

71

72

73

74

j111

Proc. Natl. Acad. Sci. U.S.A., 103, 12296–12300. Shivanyuk, A. and Rebek, J., Jr (2001) Chem. Commun., 2424–2425. (a) Shivanyuk, A. and Rebek, J., Jr (2001) Proc. Natl. Acad. Sci. U.S.A., 98, 7662– 7665; (b) Philip, I. and Kaifer, A. (2005) J. Org. Chem., 70, 1558–1564. Mecozzi, S. and Rebek, J., Jr (1998) Chem. Eur. J., 4, 1016–1022. Beyeh, N.K., Kogej, M., Åhman, A., Rissanen, K. and Schalley, C.A. (2006) Angew. Chem., 118, 5339–5342; Angew. Chem. Int. Ed., 45, 5214–5218. Dalgarno, S.J., Tucker, S.A., Bassil, D.B. and Atwood, J.L. (2005) Science, 309, 2037–2039. Kobayashi, K., Shirasaka, T., Horn, E., Furukawa, N., Yamaguchi, K. and Sakamoto, S. (2000) Chem. Commun., 41–42. (a) Ajami, D. and Rebek, J., Jr (2006) J. Am. Chem. Soc., 128, 5314–5315; (b) Ajami, D. and Rebek, J., Jr (2006) J. Am. Chem. Soc., 128, 15038–15039. Atwood, J.L., Barbour, L.J. and Jerga, A. (2001) Chem. Commun., 2376–2377.

j113

4 Cucurbit[n]urils Wei-Hao Huang, Simin Liu, and Lyle Isaacs

4.1 Introduction

This chapter focuses on the synthetic and supramolecular chemistry of the cucurbit[n]uril (CB[n]; n ¼ number of glycoluril units) family of macrocycles. Our research in this area, which began in 1998 at the University of Maryland, has been largely driven by our efforts to understand the mechanism of CB[n] formation. The chapter is therefore organized around that theme and is illustrated by subsequently developed applications of the CB[n] family. 4.1.1 Synthesis and Structure of Cucurbit[6]uril and Decamethylcucurbit[5]uril

The cucurbituril story begins in 1905, when Behrend described the condensation of glycoluril (1a) with two equivalents of formaldehyde (Scheme 4.1) under aqueous acidic conditions (HCl, 100 C) [1]. The substance formed during this reaction – referred to as Behrends polymer in the literature – was insoluble in all common solvents but could be recrystallized from hot H2SO4, which yielded a well-deﬁned substance. Although Behrend was not able to structurally characterize this substance, he did demonstrate that it forms complexes with a wide variety of species including KMnO4, AgNO3, and methylene blue. It was not until 1981 that Mock reported that the product of Behrends reaction was the macrocyclic hexameric cucurbit[6]uril comprising six equivalents of glycoluril and 12 equivalents of formaldehyde [2]. During the course of this remarkable reaction, six rings and 24 bonds are formed with complete control over the disposition of the glycoluril methine C–H groups which point outward from the central cavity. It would be nearly 10 years later that Stoddart reported that the condensation of dimethylglycoluril (1b) with formaldehyde (two equivalents) under similar conditions yielded the ﬁrst cucurbituril derivative – macrocyclic pentameric Me10CB[5] (Scheme 4.1) [3]. These two early

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

114

O

j 4 Cucurbit[n]urils OO

OO

O

O a) CH2O, HCl, NN N N N N HN NH conc. HCl heat R R N N b) H2SO4 N HN NH H2CO, reflux, NN N N N N N NN 16% O O OO O O O CB[6] (R = H) 1a R = H 1b R = Me

N N NN N N

O

OO

N N NN N N

OO

O

NN N N N N

N N NN N N N N NN NN O O OO O n O Me10CB[5] (R = Me, n = 0)

Scheme 4.1 Synthesis of CB[6] and Me10CB[5].

members of the cucurbit[n]uril family provided us, and several other groups, with numerous questions [4–6]: 1. What is the scope of glycoluril derivatives that can be used in the CB[n]-forming reaction? 2. Why does the CB[n]-forming reaction deliver hexameric CB[6] but pentameric Me10CB[5]? 3. What factors are responsible for the remarkably high yield (82%) obtained for CB[6]? 4. What is the mechanism of CB[n] formation? 4.1.2 Molecular Recognition Properties of Cucurbit[6]uril

Throughout the 1980s, Mock published a series of pioneering papers that demonstrated that CB[6] has remarkable abilities as a molecular container in aqueous solution [2,7–9]. For example, he showed that pentanediamine (2) and hexanediamine (3) bind to CB[6] as their ammonium salts with values of Ka > 106 M1 in 40% aq. HCO2H. Even more remarkable was the fact that compounds with shorter or longer spacer groups between the ammonium centers [e.g., butanediamine (4) or heptanediamine (5)] bind signiﬁcantly more weakly (Ka 104–105 M1). This high afﬁnity and selectivity can be ascribed to the juxtaposition of three distinct binding regions: (a) two symmetry equivalent ureidyl–carbonyl lined portals that have a pronounced negative electrostatic potential and interact with cations (e.g., Hþ, Mþ, or RNH3þ) by ion-dipole interactions and/or hydrogen bonds, and (b) a central hydrophobic cavity that has a slightly negative electrostatic potential which preferentially interacts with hydrophobic guests (Scheme 4.2). The high selectivity observed for 2 and 3 over 4 and 5 is due to the good match between the H3N s s sNH3 separation on the guest and the C¼O s s s O¼C separation on the host. CB[6] is also quite selective based on size and shape. For example, CB[6] binds nicely to p-methylaniline (6) but completely rejects the meta-isomer (7). Similarly, CB[6] readily encapsulates 8 which has a volume of 86 Å3 but rejects the similarly sized 9 (89 Å3). Lastly, polyammonium species like spermine (10) achieve very high afﬁnity in their binding with CB[6] (Ka ¼ 1.3 107 M1) [8]. These high-afﬁnity and highly selective binding events were utilized by Mock to achieve a 104-fold rate enhancement of a dipolar cycloaddition within the cavity [10] and in the construction of a molecular

4.2 New Members of the Cucurbit[n]uril Family

Ion-dipole intreactions

H-bonding

H3N

H N OO 3

O N

n

N N

N NN

+ CB[6] NH3

4 (n = 1) 2 (n = 2) 3 (n = 3) 5 (n = 4)

NN O

Hydrophobic effect

N

N

N

N

O

O

OO

N

NH2

H3N

O

NN N N NN

NN

NH2

j115

N

N

N

O NH3O

8

H2N

6

2

NH2

NH2

NH2

O

CB[6]•3

10 9

7

Scheme 4.2 Molecular recognition properties of CB[6].

shuttle based on changes in pH [11]. In the discussion above and in the remainder of this chapter we describe and depict the guests in the form (e.g., amine or ammonium) that was added to the aqueous solution; in nearly all cases it is the cationic ammonium form that actually binds to CB[n]. These examples convinced us that the supramolecular chemistry of CB[6], its derivatives, and it analogs would be especially well suited for the preparation of molecular machines and biomimetic systems. In 1998, therefore, we embarked on a journey to understand the mechanism of CB[6] formation, since we were conﬁdent that this would lead to the preparation of new members of the cucurbituril family with exciting new properties.

4.2 New Members of the Cucurbit[n]uril Family

This section describes the preparation of CB[n] homologs (n ¼ 5, 7, 8, 10). We begin the discussion with a mechanistic hypothesis for CB[n] formation advanced by the Day and Isaacs groups [5,12]. 4.2.1 Proposed Mechanism of Cucurbit[n]uril Formation

Although the CB[n]-forming reaction is remarkably complex, we sketched out a rough mechanism in 1998 and supplemented it along the way as new information became available [4,5,13,14]. Scheme 4.3 depicts our current understanding of the mechanism of CB[n] formation. Initially, glycoluril (1a) reacts with formaldehyde to yield methylene-bridged glycoluril dimers (11C and 11S). These two compounds are diastereomers that differ in the relative orientation of the H-atoms on the convex face of the molecule. In CB[6], however, all of the methine C–H groups project outward from the center of the macrocycle similarly to 11C. Key questions that we wished to address, therefore, were: (a) Are compounds with 11S-type stereochemistry formed? (b) Are 11S and 11C in equilibrium with each other? and (c) What is the mechanism of the 11S to 11C interconversion? These methylene-bridged

NH3

116

j 4 Cucurbit[n]urils O

HN H HN

OO

O NH H NH

O

CB [n ≥ 8]

Mineral acid, CB [n] ( n = 5-7 )

CH2O, heat

ns-CB[10]

H H

N

H H

NN

N

N

N

O NH N

HN H HN

O N

N

N O

N 11S

O

O

O N

N

NH H NH

HH N O

N O N N

O

H H

13

Template

O NH HN H+ H NH HN

HH

NN

H

H

N

Dimer Formation O

O N

N NH N O O O n-4 O

i-CB[n]

1a

N

N

N 11C

Dimer Equilibration

O

RN H RN

O N

N

O N

HH N O

N HH

N

N O

N

NR H NR

n O

12 Behrend's Polymer

Scheme 4.3 Proposed mechanism of CB[n] formation (i-CB[n], inverted cucurbit[n]uril; ns-CB[10], bis-nor-seco-cucurbit[10]uril).

glycoluril dimers can then undergo further oligomerization to yield methylenebridged glycoluril oligomers (e.g., trimer – decamer) as a mixture of S-shaped and C-shaped diastereomers (12). This mixture of oligomers has been referred to in the literature as Behrends polymer. The mixture of oligomers can then undergo S- to Cshaped isomerization, perhaps under the inﬂuence of a templating group to yield the all C-shaped oligomer (13). The ends of oligomer 13 then react with one another to enter the CB[n] manifold as a mixture of CB[n] (n ¼ 5, 6, 7, 8, 10). Through product resubmission experiments, Anthony Day demonstrated that CB[8] undergoes ring contraction to yield CB[5]–CB[7] whereas the smaller CB[n] homologs (CB[5]–CB[7]) are stable to the reaction conditions [12]. 4.2.2 Synthesis and Structure of Cucurbit[n]uril Homologs (n ¼ 5, 7, 8, 10)

The isolation of CB[n] homologs (CB[5], CB[7], and CB[8]) by the groups of Kim and Day around the turn of the millennium represented a major breakthrough in the cucurbituril ﬁeld [12,15]. Figure 4.1 shows top views of CB[5]–CB[8] (drawn to scale) which illustrate the circular shape of the cavity and the gradual progression to larger cavity volumes [15,16]. In particular, the larger homologs CB[7] (279 Å3) and CB[8] (479 Å3) with their larger cavities which parallel those of b- and g-cyclodextrin promise to greatly expand the range of applications to which CB[n] can be applied. 4.2.2.1 Reaction Conducted Under Milder Conditions To access the unknown cucurbituril homologs, Kim and Day performed the CB[n]forming reactions under milder conditions (100 C, conc. HCl; 75 C, 9M H2SO4) [12,15]. Although CB[6] is still the major component of the reaction mixture, the

4.2 New Members of the Cucurbit[n]uril Family

Figure 4.1 Molecular models of CB[5]–CB[8] and their calculated cavity volumes.

CB[n] homologs (CB[5], CB[7], CB[8], and CB[10] sCB[5]) can be readily isolated on the 10–100 g scale by straightforward – but time consuming – washing and recrystallization procedures that are detailed in Section 4.8.2. To date, CB[9] has not been isolated in pure form, although it has been detected as a minor component of the CB[n]forming reaction mixture by mass spectrometry and 13 C NMR spectroscopy [12,15]. 4.2.2.2 CB[5] Can be Released from CB[10] sCB[5] to Yield Free Cucurbit[10]uril In addition to CB[5]–CB[8], Day and coworkers were also able to isolate CB[10] as its CB[10] sCB[5] complex (Scheme 4.4) [12,17]. Day and coworkers also demonstrated that added 13 C-labeled CB[5] undergoes chemical exchange with unlabeled CB[5] in the CB[10] sCB[5] complex. This observation suggested that it should be possible to remove CB[5] and isolate free CB[10] in its uncomplexed form. From our 1 kilogram scale CB[n]-forming reaction (Section 4.8.2) we were lucky enough to isolate a signiﬁcant quantity of CB[10] sCB[5] and decided to try and remove CB[5]. We

Scheme 4.4 Sequence used to remove CB[5] from CB[10] sCB[5]. The structures of CB[10] sCB[5] [17] and CB[10] s192 [18] are based on the X-ray crystal structures, whereas those for CB[10] s19 and CB[10] are purely schematic representations.

j117

118

j 4 Cucurbit[n]urils discovered that some of the guests commonly used (e.g., 14–18) to complex to the larger CB[n] (n ¼ 7, 8) are capable of partially displacing CB[5] from the CB[10] sCB[5] complex. In CB[n]-binding studies, the free amines are commonly used when buffered D2O is the solvent, whereas the more soluble hydrochloride salts are employed when applications demand the use of D2O. Unfortunately, in many cases, all of the components (CB[5], excess guest, and CB[10] sguest) remained in solution, which complicated the isolation of the CB[10] sguest complex. We discovered that addition of excess guest 19 to CB[10] sCB[5] results in the formation of the soluble CB[10] s192 complex and precipitation of the (CB[5] s19)n supramolecular polymer (Scheme 4.4) [18]. The pure CB[10] s192 complex is isolated by ﬁltration and rotary evaporation. Removal of the less strongly bound second equivalent of 19 occurs readily by washing with MeOH, which yields CB[10] s19. The ﬁnal equivalent of 19 can be removed by heating with acetic anhydride, which transforms the ammonium groups into acetamides, followed by washing (DMSO, MeOH, H2O). Uncomplexed CB[10] has a cavity volume of approximately 870 Å3 and is poorly soluble in water and mildly acidic solution. Fortunately, its complexes have good solubility in D2O, which allowed us to study its molecular recognition properties (Section 4.3.5).

4.3 Applications of Members of the Cucurbit[n]uril Family

The early work of Mock on the recognition properties of CB[6] established that this macrocycle was amongst the tightest binding and most selective synthetic hosts in aqueous solution [2,7–9]. Subsequent work from our group demonstrated that the high afﬁnity and selectivity translates to individual host-guest complexes of CB[7] and CB[8]. Even more impressively, the selectivity of a single guest toward CB[6], CB[7], and CB[8] can be remarkably large (>104) [19]. These, and related observations [20,21], suggested to us and others that CB[n] would function as smart components in a variety of applications that are described in this section. 4.3.1 Preparation of Molecular Switches

Compounds that function as molecular switches exist in two stable forms and can be reversibly switched between those two forms by application of a given stimulus (e.g., pH change, electrochemical, photochemical, or chemical). A particularly interesting molecular switch based on the triaminoﬂuorene compound 20 was reported by Kim in 2000 (Scheme 4.5) [22]. When CB[6] is added to 203þ in pH 1.0 solution, rotaxane CB[6] s203þ is formed where the CB[6] bead resides on the longer hexylene spacer in order to maximize both ion-dipole interactions and hydrophobic binding. In this form, the anilinium N-atom is protonated and the ﬂuorene chromophore is highly ﬂuorescent. When the pH is raised (8.0), however, the anilinium N-atom is deprotonated and the CB[6] macrocycle slides over to the butylene spacer to maintain two ion-dipole interactions while sacriﬁcing some hydrophobic driving force. In this

4.3 Applications of Members of the Cucurbit[n]uril Family

Scheme 4.5 A pH-responsive molecular switch that signals by color and fluorescence changes.

new complex (CB[6] s202þ), the aniline N-atom is conjugated to the ﬂuorene ring and the complex has a violet color; the N-atom also quenches the ﬂuorescence output of the system. Upon lowering the pH, the CB[6] bead again switches back to the original location. Thus, CB[6] acts as a pH-controlled molecular switch with both UV/Vis and ﬂuorescent output. 4.3.2 Self-Assembled Dendrimers

The ability of members of the CB[n] family to bind tightly to their guests has been used by the groups of Kaifer [23,24] and Kim [25] to modify the properties of dendrons containing guests at their focal points and to bring two or more dendrons together to create a self-assembled dendrimer. For example, Kaifer synthesized 1st, 2nd, and 3rd generation Newkome-type dendrons 21a–21c containing viologen units at the focal point (Scheme 4.6) [23]. UV/Vis titrations allowed Kaifer to demonstrate that the association constant for the formation of the binary complex CB[8] s21 decreases as the generation of the dendrimer increases. The electrochemical behavior was even more interesting. Upon reduction the viologen is transformed from its dicationic to its monocationic form, which subsequently induces dimerization inside CB[8] yielding ternary complex CB[8] s212 as evidenced by changes in the UV/ Vis spectrum. Interestingly, CB[8] s22 did not form CB[8] s22 upon reduction, presumably because of unfavorable electrostatic interactions. The dramatic change in size of the ternary complex was also determined by measurement of its diffusion constant by pulsed ﬁeld gradient NMR techniques. No dimerization is observed in the absence of CB[8]. This beautiful example demonstrates proof-of-concept for redox-switchable dendrimer self-assembly. 4.3.3 Preparation of Molecular Machines

The Kim group has capitalized on the ability of CB[8] to simultaneously bind two aromatic rings in the construction of a variety of molecular machines [26,27]. A

j119

120

j 4 Cucurbit[n]urils

Scheme 4.6 Electrochemical control of dendrimer assembly.

recent and particularly elegant example is the construction of a molecular loop lock (Scheme 4.7) [27]. They found that the interaction of CB[8] with 233þ leads to the folded conformation of CB[8] s233þ (Scheme 4.7), which beneﬁts from intramolecular charge transfer interaction between the electron-rich naphthalene and electronpoor viologen units. The addition of one equivalent of methylviologen (242þ) does not result in a change in conformation of CB[8] s233þ, presumably because 242þ binds less strongly than the intramolecularly held viologen unit. Very interestingly, however, upon reduction of the solution of the mixture of CB[8] s233þ and 242þ with Na2S2O4, the equilibrium shifts toward the unfolded ternary complex CB[8] s 232þ s24þ. This ternary complex beneﬁts from the enhanced p-p interactions, as evidenced by the appearance of new bands at 368, 550, and 890 nm in the UV/Vis spectrum of the viologen radical-cation dimer. The system is fully reversible between the CB[8] s232þ s24þ and CB[8] s233þ states by exposing the system to molecular

Scheme 4.7 A molecular loop lock that responds to chemical and redox stimuli.

4.3 Applications of Members of the Cucurbit[n]uril Family

oxygen (O2). This system may be regarded as a safeguarded molecular loop lock in that it requires not only the addition of 242þ, but also the application of a redox stimulus (Na2S2O4 or O2) to open or close the lock. 4.3.4 Preparation of Complex Self-Sorting Systems

Mock has shown that CB[6] undergoes tight and selective binding with cationic guest molecules in aqueous solution [8]. Our investigations of complex (e.g., 12-component) self-sorting mixtures which utilized CB[6] and CB[8] as components suggested that the binding strengths (Ka) and selectivities might be even higher than previously appreciated [20]. To provide quantitative evidence for this hunch, we measured the binding constants for CB[6], CB[7], and CB[8] toward a variety of guests by 1 H NMR competition experiments [19]. We found that values of Ka ranged between 102–1012 M1 and moreover that selectivities could be exceedingly large (>103) in response to small structural changes. For example, CB[8] binds 19 3250-fold tighter than CB[7] and induces a folding process at the same time (Scheme 4.8). Similarly, CB[7] binds 14 5150-fold tighter than CB[8] because of an excellent size, shape, and electrostatic match between host and guest. In addition to the high afﬁnity and selectivity, members of the CB[n] family can exhibit unusual dynamics of complex association and dissociation [9,21,28]. We used the high binding afﬁnities, high selectivities, and unusual dynamics to construct a system that undergoes well-deﬁned transitions from free components, to one set of aggregates under kinetic control, to a second set of aggregates under thermodynamic control (Scheme 4.9). For example, when a solution of CB[6] and CB[7] is mixed

Scheme 4.8 High-selectivity binding within CB[n].

j121

122

j 4 Cucurbit[n]urils

Scheme 4.9 Kinetic versus thermodynamic self-sorting using two-faced guests.

with a second solution containing guests 25 and 26 the components initially associate to form CB[6] s26 and CB[7] s25. In this process, 25 undergoes fast kinetically controlled association with CB[7] in preference to CB[6] to form CB[7] s25. Subsequently, 26 – which we refer to as a two-faced guest since it contains two binding epitopes – uses its slimmer alkylammonium binding epitope to associate with CB[6]. This kinetically controlled self-sorting state comprising CB[6] s26 and CB[7] s25 slowly transforms over the course of 56 days into the thermodynamic self-sorting state comprising CB[6] s25 and CB[7] s26. NH2(CH2)3CH3 Cl

Cl

Cl NH3

H3N 25

26

4.3.5 Allosteric Control of the Conformation of a Calix[4]arene Inside CB[10]

After we synthesized CB[10] we wanted to investigate its abilities as a host molecule. We found that CB[10] forms host-guest complexes with many of the guests typically used with the smaller members of the CB[n] family (e.g., ammonium salts of alkyl and arylamines, adamantanes, viologens, and ferrocene derivatives like 14–19). We discovered that CB[10] was also capable of forming a complex with cationic calix [4]arene 27, which itself usually functions as a host molecule in aqueous solution (Scheme 4.10) [18]. Interestingly, when 27 is incorporated inside CB[10], two sets of resonances are observed which were in slow exchange on the NMR

4.3 Applications of Members of the Cucurbit[n]uril Family

Scheme 4.10 Allosteric control of the macromolecular conformation inside CB[10].

time scale. The major set of resonances corresponds to the CB[10] s1,3-alt-27 conformer, whereas the minor set of resonances corresponds to a dynamic equilibrium between CB[10] s1,2-alt-27, CB[10] spartial cone-27, and CB[10] scone-27 conformations. Scheme 4.10 shows the MMFF calculated geometry of the CB[10] s1,3-alt-27 conformation, which shows a good match between the circular shape of 1,3-alt-27 and the cavity of CB[10]. We wondered whether it would be possible to stabilize one of the other three conformers of the macromolecular CB[10] s27 complex by the addition of a guest molecule. After some experimentation, we found that adamantane carboxylic acid 28 is capable of shifting the equilibrium toward the CB[10] scone-27 s28 complex. In this ternary complex, the calixarene 27 assumes the cone-conformation in order to maximize non-covalent interactions with 28. The cone-27 s28 complex in turn binds within CB[10]. In the MMFF-computed conformation of CB[10] scone-27 s28, we once again saw a good match between the overall size, shape, and electrostatic complementarity between the cone-27 s28 complex and the cavity of CB[10]. To complete the cycle of allosteric control of conformation within CB[10], we added CB[7], which binds tightly to 28 forming CB[7] s28 and switches the conformation of calix[4]arene 27 back to the original mixture of CB[10] s27 conformers. This result demonstrated that CB[10] has great potential in controlling the conformation of non-natural macromolecular complexes. CH2NMe2

CH2

27

OH

CO2H

4

28

4.3.6 As a Carrier of Anti-Cancer Agents

One of the potential uses of any molecular container compound is in the formulation of active pharmaceutical agents. Accordingly, the groups of Kim [29] and Day [30] have demonstrated the ability of CB[7] and CB[8] to modify the properties of Pt-based anticancer agents (Scheme 4.11). For example, Kims group reported that CB[7] forms a 1:1 complex with oxaliplatin (CB[7] s29). They found that encapsulation of the drug

j123

124

j 4 Cucurbit[n]urils

Scheme 4.11 Complexation of oxaliplatin 29 within CB[7].

within CB[7] resulted in a dramatic increase in its lifetime (e.g., 6 h versus >1 year). Furthermore, CB[7] s29 showed markedly decreased reactivity toward L-methionine as a protein model relative to its reactivity toward guanosine as a DNA model. This result suggested that encapsulation of 29 within CB[7] may reduce unwanted side effects caused by protein binding of the drug.

4.4 Experimental Support for the Proposed Mechanism of CB[n] Formation

Although none of the applications of the CB[n] family described in Section 4.3 were known in 1998 when we began our work in the area, they provide post-facto support for our assertion that a ﬁrm knowledge of the mechanism of CB[n] formation would be important. This section describes our work with S- and C-shaped methylenebridged glycoluril dimers that provided insight into the early steps of the mechanism of CB[n] formation. 4.4.1 S-shaped and C-shaped Methylene-bridged Glycoluril Dimers

The high level of complexity of the CB[n]-forming reaction led us to initially study S-shaped and C-shaped methylene-bridged glycoluril dimers (11C and 11S, Scheme 4.3). This section details the lessons learned from those model studies. 4.4.1.1 Synthesis of Methylene-bridged Glycoluril Dimers Although we wanted to study methylene-bridged glycoluril dimers 11C and 11S, we were mindful of the fact that the presence of four free ureidyl NH groups and the absence of solubilizing groups on the convex face of the glycoluril skeleton might lead to poor solubility characteristics. Thus, in accord with the precedent of the Nolte and Rebek groups, we decided to target compounds whose NH-groups were protected by o-xylylene groups and that contain solubilizing substituents on their convex face [31]. Although we have investigated this reaction in detail, the bare essence of the process is presented in Scheme 4.12. In brief, we discovered three pathways to methylene-bridged glycoluril dimers that each occured in reﬂuxing dichloroethane containing p-toluenesulfonic acid with azeotropic removal of water.

4.4 Experimental Support for the Proposed Mechanism of CB[n] Formation

O

Br N

NH a) R NH

R N Br

O

Br N

N

N

N

Br

RO

N

b)

N

N

N N NO2 (±)-34

N

33C

N

N

N

N

N

N

N

RR N 33S

R

O

N

O N

OMe

N

RR N

OMe (±)-36C (31%)

O

R N

O

N O

10%

O

OMe

N

N

+

O

N NO2

R

R

O

R

N

N

O

R

O

N

N

N

N

O N

N

N

31S

O

92%

O

Br

N RR

N

RR

NO2

RO

O

N

N

N

N

N

N

N

Br

+

O

OMe

R

N R

N

N

O

O

R

b)

Br

R

OMe

O + O

MeO

N

N

NH R NH

R

47%

O

RR

O O

+

R

N

N

OMe

35

N

O

Br

O

R

O 32

MeO

N

O N

N

N

31C

O

O

R

Br

RR

R

O 30

N

O

j125

OMe (±)-36S (64%)

Scheme 4.12 Synthesis of S- and C-shaped methylene-bridged glycoluril dimers. R ¼ CO2Et. Conditions: (a) PTSA, (CH2O)n, ClCH2CH2Cl, reflux; (b) PTSA, ClCH2CH2Cl, reflux (n.d. ¼ not detected).

The ﬁrst method involves the homodimerization of glycoluril NH-compounds (e.g., 30) with CH2O and delivers both 31C and 31S. The second involves the homodimerization of glycoluril cyclic ethers (e.g., 32), which occurs with formal extrusion of formaldehyde and selectively delivers 33C (33S not detected). The third involves heterodimerization of a glycoluril NH-compound [e.g., ()-34] with a glycoluril cyclic ether (e.g., 35), which delivers a mixture of ()-36C and ()-36S. In this example of heterodimerization, the unsymmetrical substitution on the aromatic ring renders ()-34, ()-36C, and ()-36S chiral and racemic. All three reactions occur in good to high yield and generally deliver the C-shaped diastereomers selectively. All three reactions are tolerant of substitution on the convex face of the glycoluril units (e.g., R ¼ Ph, (CH2)4, CO2Et, CO2H, imide) and substitution on the o-xylylene rings (e.g., OMe, CH3, Br, NO2, F). Interestingly, the heterodimerization reaction allows the selective preparation of C-shaped methylene-bridged glycoluril dimers (e.g., 36C) containing differentially functionalized aromatic rings. These three methods not only allow us to prepare a wide variety of methylenebridged glycoluril dimers for studies of self-assembly; they also suggest methods for

n.d.

Br

126

j 4 Cucurbit[n]urils the tailor-made synthesis of partially substituted CB[n] from appropriate glycoluril building blocks (Section 4.4.2). 4.4.1.2 S- to C-shaped Isomerization of Methylene-bridged Glycoluril Dimers In our synthesis of methylene-bridged glycoluril dimers (Section 4.1), we were somewhat surprised that only very small amounts of the S-shaped diastereomers were observed in most cases. To determine whether this diastereoselectivity was due to thermodynamic or kinetic preferences we isolated diastereomerically pure samples of six C-shaped and S-shaped compounds and separately resubmitted them to the reaction conditions. Scheme 4.13 shows the separate isomerization of 37C and 37S. We found that both the C-shaped and S-shaped forms delivered an equilibrium mixture of S-and C-shaped forms in which the C-shaped form predominates (95:5) for a variety of compounds. This result established that the C-shaped form is thermodynamically more stable than the S-shaped form by 2 kcal mol1 and provided a rationale for the high yield observed in the formation of CB[6], which presumably required all C-shaped oligomers. To delve deeper into the origin of this thermodynamic preference we performed the equilibration of 37C and 37S in a variety of solvents (e.g., CHCl3, CCl4, C6F6, THF, CH3CN, ClCH2CH2Cl, CH3NO2, MeOCH2CH2OMe). The C-shaped to S-shaped ratio remained high (>90:10) across this series, which indicated that solvation was not a key factor determining the thermodynamic preference for the C-shaped form. We also ruled out the possibility that PTSA acted as a template by selectively engaging in p-p interactions with the cavity of the C-shaped form. Currently, we believe that the C-shaped form is more stable because of the conformational preference of the newly formed 8-membered ring for the crown conformation [32]. 4.4.1.3 Mechanism of S- to C-shaped Isomerization The selective heterodimerization of glycoluril NH-compounds and glycoluril cyclic ethers (Section 4.4.1.1) and the general preference for the C-shaped diastereomers suggested that the mechanism of S-shaped to C-shaped equilibrium might proceed by an intramolecular rather than an intermolecular process under anhydrous acidic

Br

R

Br

R N

N

N

R

N

N

Br

O

O

R

O Br

N

N

R

Br 37S (S-shaped)

reflux R = CO2Et

N

N

N

PTSA, ClCH2CH2Cl

N

N

O

R

Br

O

O %C

N

N O O

R

100

N

Equilibrium C:S ratio

Br

N R

0

Br 37C (C-shaped)

Scheme 4.13 Equilibration of 37C and 37S.

Time

4.4 Experimental Support for the Proposed Mechanism of CB[n] Formation O

OMe N

O N

R

N

OMe N

RR N

NO2

N

PTSA R

N

N

O N

O N

R

N

N NO2

N

N

N

N

N O

N

41

N

R

NO2

O N N

N

N

N

R

N

N

N

N

N

N

RR N

N O

R

H+

O

N

N

N

OH

O

O •

N

N

R

N

N

N

N

O

R

OH 43

N N

N

N

N

R N N

42

R

44

O

RR N

N

•

O

R

R

NO2

O O

R

•

OMe

N

39C

O

•

•

OMe

RR

O

N

N

O

RR N

N

O

N

N

N

O (±)-38C O

N

•

NO2

N RR

OH N

O

ClCH2ClCH2

N

N H+

R N

N

N

R

O •

N

R

OMe

RR

R

40

OMe

N

N

O N

N

NO2 PTSA

O

O •

NO2

R

39S

O

NO2

N

RR

O

R

ClCH2ClCH2

O (±)-38S O

OMe

O

OMe

R

j127

•

O O

Scheme 4.14 Mechanism of diastereoselective intramolecular isomerization of S- to C-shaped compounds.

conditions (e.g., PTSA, ClCH2CH2Cl, reﬂux). To test this hypothesis, we designed a double-labeling experiment that utilized ()-38 and 39 (Scheme 4.14). In the equilibration of chiral but racemic ()-38S, for example, one might expect to obtain either ()-38C or 39, or potentially a mixture of both [5]. Similarly, the equilibration of 39S might deliver either ()-38C or 39, or both. In the experiment, we observe that ()-38S cleanly delivers the chiral but racemic ()-38C whereas the achiral meso compound 39S cleanly delivers the achiral meso compound 39C. This result indicates that the transformation of S- to C-shaped methylene-bridged glycoluril dimers is an intramolecular process under anhydrous acidic conditions that proceeds via retention of conﬁguration of the two glycoluril halves. Scheme 4.14 also shows a proposed mechanism that tracks the position of the labels (green dots). Initially, 40

•

128

j 4 Cucurbit[n]urils undergoes protonation and ring opening to yield iminium ion 41. Iminium ion 41 can undergo ring closure by nucleophilic attack of the glycoluril N-atom to yield spirocyclic intermediate 42, which can fragment in the opposite direction to yield iminium ion 43. Subsequent ring closure of 43 followed by deprotonation gives 44, which rationalizes the experimental results. 4.4.1.4 Implications for the Synthesis of Cucurbit[n]uril Analogs and Derivatives The results presented in Sections 4.4.1.1–4.4.1.3 provide insight into the mechanism of CB[n] formation. For example, the isolation of the S-shaped diastereomers provides strong evidence that both 11S and 11C form in the initial stages of the CB[n]-forming reaction. The strong preference for the C-shaped diastereomers (>90:10) provides a rationale for the high yield obtained in the synthesis of CB[6] [33]. A simple calculation suggests that in a worst case scenario 53% (0.96) of the linear glycoluril hexamers would exist as the all C-shaped diastereomer required to form CB[6]. In addition, the intramolecular nature of the S-shaped to C-shaped equilibration suggests that the length of the growing methylene-bridged glycoluril oligomer chain (13, Scheme 4.3) determines the size (e.g., n) of the CB[n] formed. When considered together with highly selective heterodimerization reactions, these results further suggest that the pattern of substituents on the surface of the growing methylene-bridged glycoluril oligomer will be preserved in the CB[n] derivatives that are formed. As a speciﬁc example, we predicted that heating a combination of 1a and bis(cyclic ether) 45a under anhydrous acidic conditions (e.g., PTSA or MeSO3H) would deliver the hexa-substituted CB[6] derivative 46a (Scheme 4.15) [5]. Although this reaction did not yield the desired product 46a (R ¼ CO2Et) because of substituent effects, the predictions made above were subsequently borne out in the synthesis of CB[n] derivatives (e.g., 46b, R ¼ Me) and CB[n] analogs (Sections 4.4.2 and 4.4.3). 4.4.2 Building-Block Approach to Cucurbit [n]uril Analogs

We hypothesized that the reaction depicted in Scheme 4.15 did not proceed smoothly in practice because the CO2Et-groups on the convex face of the glycoluril reduce its reactivity (sterically and electronically) relative to the unsubstituted glycoluril 1a. To

N NH H+ O R N NH

HN H HN O 1a

N

Anh. Acid R O

N O

R R

N

N N

N NN

NN O

45a R = CO2Et 45b R = Me

Scheme 4.15 Predicted synthesis of CB[n] derivatives.

OO

OO

O

O

O

N

N

N O

N

NN

O

N N NN

R R

R R NN

N

N

N

O OO O 46a (EtO2C)6CB[6] 46b Me6CB[6]

N

4.4 Experimental Support for the Proposed Mechanism of CB[n] Formation

O

O N

N

N

OR

N

RR N

N

NH NH O

O

48

47

O N

N

R

N

N R

O

O N

N

N

O

N

N

O

N

O

N

O

N

O

N

O

N

N

N

N

N

N

RR N

N

RO

O

49

O

47, MeSO3H, 80 ºC R

N

N

O 50 (78%)

N O

N

R

O

N O

N RR

O

R

N

O

NN

R

O

N O

N

O

N

N

N OR

O

47, MeSO3H, 80 ºC R

O

O

HN HN

RO N

N

O

O

O

j129

R

N

N

N

R

O

NN

O

O

N O

N O

R

R

N

N

N R

R

R R R

N O

R

N NN O O H (±)-51 (67%)

O

N

O

N

O

N

O

O

N

R

N

N H N

O

N

N

N

N

Scheme 4.16 Synthesis of CB[n] analogs. R ¼ CO2Et.

circumvent this issue we decided to search for a glycoluril analog that might participate efﬁciently in heterodimerization reactions with compounds like 45 to yield CB[n] analogs. With a little bit of luck we discovered that phthalhydrazides are superior nucleophiles – presumably because of the a-effect – that undergo selective heterodimerization with glycoluril cyclic ethers under anhydrous acidic conditions [34–36]. Speciﬁcally, bis(phthalhydrazide) 47 was found to undergo separate macrocyclization with dimeric 48 and trimeric 49 to yield CB[n] analogs (50 and ()-51) in remarkably high yield (Scheme 4.16). These macrocycles incorporate the UV/Vis, ﬂuorescent, and electrochemically active bis(phthalhydrazide) walls, and the carboxylic acid version of 50 can be used as a ﬂuorescent sensor in aqueous solution [36]. Compound ()-51 was particularly intriguing since the presence of two free ureidyl NH-groups and the twisted connection between the tips of the two glycoluril trimer building blocks 49 imparts C2-symmetry to the macrocycle and renders ()-51 chiral and racemic. Compound ()-51 is the only chiral member of the CB[n] family known to date. 4.4.3 Building-Block Approach to Cucurbit [n]uril Derivatives

Concurrent with the synthesis of the CB[n] analogs, the groups of Day and Tao were also pursuing a building-block approach toward CB[n] derivatives [37,38]. For example, Day was able to isolate Me6CB[6] (46b) from the reaction of 1a and 45b using aqueous acidic reaction conditions (Scheme 4.15). An elegant example of a buildingblock approach toward CB[n] derivatives was realized by reacting the unsubstituted methylene-bridged glycoluril dimer 52 with cyclic ether 45b, which delivered the tetra-substituted CB[6] derivative Me4CB[6] in 30% yield (Scheme 4.17). Quite interestingly, Me4CB[6] has an ellipsoidal shape which binds aromatic compounds with a common orientation of the long axes of both host and guest [38].

O

N

O

N

N

N R

R

130

j 4 Cucurbit[n]urils

O HN H HN

N

N

NH H NH

HH N O

O

O

N 52

O

N +

O N

O Me

Me O N

N O

45b

HCl / H2O 100 ºC

OO

OO

O

N N

Me Me

NN NN N N NN NH N H H Me H H Me H H H N N N N NN N N NN N N

O

O

OO O Me4CB[6]

O

Scheme 4.17 Synthesis of Me4CB[6].

4.4.4 Identification and Isolation of Inverted Cucurbit[n]urils (n ¼ 6, 7)

The preferred – but not exclusive – formation of C-shaped methylene-bridged glycoluril dimers provided a rationale for the observed high yields in CB[n]-forming reactions. Conversely, those same experiments can also be interpreted as suggesting the possibility of diastereomeric CB[n]. Such diastereomeric CB[n], in which one or more pairs of glycoluril H-atoms point into the central CB[n] cavity, are referred to as inverted CB[n] (i-CB[n]). The ﬁrst members of the i-CB[n] family were discovered by the groups of Isaacs and Kim and investigated jointly (Scheme 4.18) [13]. i-CB[6] and i-CB[7] are formed in small amounts in the CB[n]-forming reaction and can be detected in the 1 H NMR spectrum of the mixture by diagnostic resonances for the inverted H-atoms (5 ppm, see below). Pure i-CB[6] and i-CB[7] could be isolated in low yield (2.0 and 0.4%) by a combination of fractional recrystallization and gel permeation chromatography. Scheme 4.18 also shows the X-ray crystal structures of i-CB[6] and i-CB[7] and their electrostatic potential energy surfaces. The

Scheme 4.18 Chemical structures, X-ray crystal structures, and electrostatic potential maps for i-CB[6] and i-CB[7].

4.5 Direct Functionalization of Cucurbit[n]urils

j131

consequences of the inverted glycoluril ring are reduced cavity volumes, ﬂattened ellipsoidal-shaped cavities, more positive intra-cavity electrostatic potentials, and more open C¼O lined portals. In accord with these considerations, i-CB[6] and i-CB[7] preferentially bind ellipsoidal and electrostatically negative aromatic guests like p-xylylene diamine (53) and ferrocene derivative 54. Overall, the inverted glycoluril modulates the guest binding afﬁnity and rates of dissociation in a way that might be useful in the creation of functional CB[n]-based systems.

4.5 Direct Functionalization of Cucurbit[n]urils

One of the major challenges in CB[n] chemistry is the poor solubility of CB[n] in neutral water and far worse solubility in organic solvents. A potential solution to this problem that has been pursued by several groups involves the preparation of CB[n] derivatives or analogs that contain solubility-enhancing functional groups on the convex face of the macrocycle. This strategy is easier to propose than to execute, and the sections above detail some of the challenges of using glycoluril derivatives in CB[n]-forming reactions. Perhaps the most onerous issue to be addressed in such a route is the modiﬁcation of the separation scheme for each new CB[n] derivative. An alternative approach to the preparation of CB[n] derivatives is the direct functionalization of pure, unfunctionalized CB[n]. 4.5.1 Perhydroxylation and Further Derivatization of CB[5]–CB[8]

Kim and coworkers addressed a major issue in CB[n] chemistry with their discovery of a method for direct functionalization [39]. They found that heating CB[5], CB[6], CB[7], or CB[8] with K2S2O8 in water resulted in the perhydroxylation of CB[n] (n ¼ 5–8) yielding (HO)2nCB[n] in modest to low yield (Scheme 4.19). Furthermore, the (HO)2nCB[n] had excellent solubility in DMSO and DMF, which allowed for subsequent functionalization. For example, (HO)2nCB[n] (n ¼ 5, 6) underwent reaction with NaH and allyl bromide in DMSO to yield (H2C¼CHCH2O)2nCB[n] in good yields. (H2C¼CHCH2O)2nCB[n] (n ¼ 5, 6) undergo photochemical addition

CB[n]

a)

(HO)2nCB[n]

(HO)10CB[5] (42%) (HO)12CB[6] (45%) (HO)14CB[7] (5%) (HO)16CB[8] (4%)

O OO OO O n=6 N N NN N N N N N N N N R n = 5, 6 R R R R R R R R R R R n = 5, 6 N N N NN N N N N NN N n=6 O O OO O n-5 O R2nCB[n]

Scheme 4.19 Direct functionalization of CB[n]. Conditions: (a) K2S2O8, H2O, 85 C; (b) NaH, DMSO, allyl bromide; (c) CH3(CH2)7SH, hn; (d) PEG-thiol, hn.

R = OH

b)

R = OCH2CH=CH2 R = O(CH2)3S(CH2)7CH3

c)

R = O(CH2)3S(CH2CH2O)3CH3

d)

132

j 4 Cucurbit[n]urils of thiols (e.g., CH3(CH2)7SH or CH3[O(CH2)2]3SH) to yield the highly lipophilic {CH3(CH2)7S(CH2)3}2nCB[n] and {CH3[O(CH2)2]3S(CH2)3}2nCB[n]. 4.5.2 Multivalent Binding of Sugar-Decorated Vesicles to Lectins

The study of vesicles is an area of wide interest largely because of their potential use in the areas of drug and gene delivery, biomimetic systems, and in the preparation of nanostructured materials. Kims group discovered that the amphiphilic CB[6] derivative {CH3[O(CH2)2]3S(CH2)3}2nCB[6] formed vesicles of 170 50 nm diameter as evidenced by transmission electron microscopy (55, Scheme 4.20) [40]. When they were prepared in a solution containing sulforhodamine G and puriﬁed by gel permeation chromatography, the vesicles showed bright ﬂuorescence, which demonstrates their low permeability. The efﬁcient non-covalent derivatization of the exterior of vesicles 55 was possible simply by the addition of the spermineﬂuorescein conjugate 56 to yield 55 s56n. Conjugate 56 was bound strongly to the cavities of the CB[6] derivatives on the surface of the vesicles by the hydrophobic effect and ion-dipole interactions as evidenced by 1 H NMR spectroscopy and confocal microscopy. This vesicle scaffold could also be used for the multivalent display of a-mannose groups tagged with a spermine tail. These a-mannose-derived vesicles were capable of aggregating concanavalin A by biospeciﬁc interactions between the sugar and the lectin. The potential for the introduction of multiple tags onto a single vesicle suggested a route toward the targeted delivery of pharmaceutical agents. 4.5.3 Cucurbit [n]uril-based Artificial Ion Channels

One of the major issues that has faced the development of applications of the CB[n] family – which has partially faded with Kims direct functionalization route – is poor solubility in organic solvents. To improve the solubility in organic solvents and in

Scheme 4.20 Decoration of supramolecular vesicles with fluorescent labels.

4.6 Nor-Seco-Cucurbit[10]uril

j133

O N

N CH2 R N CH2

R H+, Li+, Na+, K+, Rb+, Cs+

N O

n = 5, 6

(H17C8SC3H6O)12CB[6] (H17C8SC3H6O)10CB[5] Scheme 4.21 Preparation of artificial ion-channels with CB[5] and CB[6] derivatives.

lipid membranes, Kim and coworkers attached long alkyl chains to the outside of CB[5] and CB[6] as described above [39]. They found that {CH3(CH2)7S(CH2)3}2n CB[n] (n ¼ 5, 6) could be incorporated into large unilamellar vesicles (LUVs) comprising egg yolk L-a-phosphatidyl-choline (EYPC), cholesterol, and dicetyl phosphate (Scheme 4.21) [41]. The ﬂux of protons and alkali metal cations across the membrane could be quantiﬁed by following the ﬂuorescence intensity (I460/I403) of 8hydroxypyrene-1,3,6-trisulfonate (HPTS), which functions as a pH-sensitive dye. Interestingly, the order of transport activity across LUVs containing {CH3(CH2)7S(CH2)3}2nCB[n] was Liþ > Csþ Rbþ > Kþ > Naþ, which is opposite to the binding afﬁnity of CB[6] toward alkali metal cations. Somewhat surprisingly, planar bilayer conductance measurements show that the ion transport occurs by a channel rather than a carrier mechanism. The ion-channel properties of {CH3(CH2)7S(CH2)3}12CB[6] could be switched off by the addition of acetylcholine (Me3NCH2CH2OAc, 57) which binds in the cavity of the CB[6] derivative. The CB[6] derivative {CH3(CH2)7S(CH2)3}2nCB[n] permits remarkably fast transport of Csþ across the membrane (5 pA 3 · 107 ions s1), which is comparable to that observed for gramicidin. By using synthetic chemistry to tailor the structure and solubility of CB[n] and its derivatives, it is possible to expand the range of applications to which the CB[n] family can be applied!

4.6 Nor-Seco-Cucurbit[10]uril

Our model studies using the methylene-bridged glycoluril dimers allowed us to shed light on the initial steps in the CB[n]-forming reaction (see Section 4.4). For example, we provided evidence that both C-shaped and S-shaped compounds were formed initially, determined that the C-shaped forms are more stable than the S-shaped forms by approximately 2 kcal mol1, and showed that the S-shaped to C-shaped isomerization reaction is an intramolecular process. To get information about the latter stages of the CB[n]-forming reaction – the elongation to trimer, tetramer, pentamer, hexamer, and longer oligomers and their subsequent macrocyclization reactions – we decided to decrease the number of equivalents of CH2O used in the

134

j 4 Cucurbit[n]urils reaction. We hypothesized that reducing the number of equivalents of CH2O would increase the stability of partially formed CB[n] macrocycles and thereby provide information on plausible intermediates in the CB[n]-forming reaction. 4.6.1 Detection and Isolation of Nor-Seco-Cucurbit[10]uril

When we reacted glycoluril (one equiv) with formaldehyde (1.67 equiv) in conc. HCl at 50 C we observed the formation of a precipitate after three days (Figure 4.2). The precipitate (ns-CB[10]) was isolated by centrifugation and puriﬁed by washing with 1:1 H2O:HCl and MeOH. Characterization of this compound by spectroscopic means alone proved challenging; luckily, we were able to obtain crystals of ns-CB[10] as its ns-CB[10] s582 complex. Figure 4.2 shows a cross-eyed stereoview of the structure of ns-CB[10] s582 in the crystal. Several features of structure are noteworthy: (a) the lack of two CH2-groups, which reveals the presence of four free ureidyl NH groups (similarly to ()-51), (b) the presence of two symmetry-equivalent cavities, and (c) the lack of vertical registration of the two cavities. The isolation of ns-CB[10] provides evidence for the existence of pentamer fragments in the CB[n]-forming reaction and further insight into the condensation pathways that are possible between methylenebridged glycoluril oligomers [14]. 4.6.2 Molecular Recognition Properties of Nor-Seco-Cucurbit[10]uril

After we isolated ns-CB[10] we decided to study its molecular recognition behavior. Initially, we sought to deﬁne the functional group preferences of ns-CB[10] and its

Figure 4.2 Chemical structure of ns-CB[10], some of its guests, and cross-eyed stereoview of the X-ray crystal structure of ns-CB[10] s582.

4.7 Summary and Conclusions

R

H3N NH3

H3N

Fe

NH3 +

+

Fe

ns-CB[10]

53 2Cl-

R

+ R

H3N H3N

j135

Fe

NH3 61 2Cl-

NH3

R = CH2NH3

Scheme 4.22 Homotropic allostery inside ns-CB[10].

effective cavity size. We found that ns-CB[10] is capable of binding with a wide range of cationic guests that typically interact with CB[n] (e.g., 3, 14–18, 25, 53, 54, 58, 59) as well as dyes, peptides, lipids, and electrochemically active substances. In these cases, ns-CB[10] forms ternary – rather than binary – complexes with these guests (i.e., ns-CB[10] sGuest2) with 1:2 host:guest stoichiometry. Larger guests (19 and 60) ﬁll both cavities simultaneously with formation of binary complexes (e.g., ns-CB[10] s Guest). The wide range of sizes of guests that form ternary complexes with ns-CB[10] suggests a high level of structural responsiveness of this host. Indeed, molecular modeling (MMFF) showed that the central non-bonded CH2. . .CH2 distance changes from 5.50 Å (ns-CB[10] s532) to 9.27 Å (ns-CB[10] s612) as the size of the guest increases [14]. The overall cavity volume undergoes a concomitant increase from 450 Å3 to 740 Å3 as the guest changes from 53 to 61 (Scheme 4.22). We wondered whether the responsiveness of the host to the size and shape of the guest would result in an allosteric response. To test for the presence of a guest-induced homotropic allostery, we combined ns-CB[10] with mixtures of guests composed of one small (e.g., 53 or 3) and one large (e.g., 61 or 59) guest (Scheme 4.22). In such cases, we observed exclusive formation of ns-CB[10] s532 and ns-CB[10] s612, with no evidence of the cross-over hetero-ternary complex ns-CB[10] s53 s61. In contrast, when mixtures of similarly sized guests (e.g., 59 and 61) were used, a mixture of homomeric and heteromeric ternary complexes was formed. This result demonstrates that ns-CB[10] is capable of distinguishing between molecules with modest to large differences in size.

4.7 Summary and Conclusions

The ﬁrst synthesis of CB[6] by Behrend in 1905, its structural elucidation by Mock and coworkers in 1981, and subsequent studies by Mock and coworkers throughout the 1980s demonstrated the signiﬁcant potential of CB[6] in studies of molecular recognition in aqueous acidic solution. The ﬁeld has taken rapid steps forward since 2000 with the discovery of the CB[n] homologs CB[5], CB[7], and CB[8] by the groups of Kim and Day, whose cavity volumes span and exceed those available with a-, b-, and g-cyclodextrin. For example, the smaller CB[5] readily encapsulates gases [42],

R

136

j 4 Cucurbit[n]urils whereas the larger CB[n] homologs encapsulate species with signiﬁcant biological and chemical potential (e.g., anticancer agents, dyes, peptides, viologens, and ferrocenes). When we began our work in this area in 1998 we decided to tackle the issue of preparing CB[n] derivatives that contained solubilizing groups on the exterior of the macrocycle that would extend the range of applications to which CB[n] could be applied. This chapter has described the mechanistic approach, pursued by our group, that initially targeted the preparation of methylene-bridged glycoluril dimers as the fundamental building blocks of the CB[n] family. These model studies allowed us to provide evidence for some of the key early steps in the CB[n] formation mechanism, namely: (a) the presence of both S- and C-shaped diastereomers, (b) the presence of a thermodynamic driving force in favor of the C-shaped form (2 kcal mol1), and (c) the intramolecular nature of the S- to C-shaped equilibrium under anhydrous acidic conditions. These observations led us to propose a building-block strategy toward the preparation of CB[n] compounds [4,5] which was subsequently demonstrated experimentally in the synthesis of CB[n] derivatives and analogs [34,37,38]. Kims synthetic advance in developing the direct per-hydroxylation of CB[5]–CB[8] allowed his group to prepare a wide range of per-functionalized CB[5] and CB[6] derivatives and to use them in new applications such as synthetic ion-channels and supramolecular vesicles [39–41]. More recently, in collaboration with Kims group, we isolated inverted cucurbiturils (i-CB[6] and i-CB[7]), which are kinetic intermediates along the pathway toward CB[n], whose transformation into CB[5], CB[6], and CB[7] was demonstrated by product resubmission experiments [13]. These key experiments provided evidence on the fate of some of the ﬁnal intermediates along the mechanistic pathway toward CB[n] and stimulated us to further probe the latter stages of the mechanism of CB[n] formation. Our key realization was that CB[n]-forming reaction mixtures containing a deﬁciency of formaldehyde would be unable to proceed to completion and therefore deliver intermediates as ﬁnal products. Based on this idea, we were able to isolate ns-CB[10], which is formally derived from CB[10] by the removal of two CH2 groups with subsequent bond reorganization. ns-CB[10] has displayed recognition properties including those reminiscent of biological systems such as the stabilization of ternary complexes and homotropic allostery based on guest size. If the CB[n] family is to supplant the cyclodextrins as the platform of choice for basic and applied studies of molecular recognition, then a series of additional developments must take place, many of which involve the development of new synthetic procedures. First, improved methods for the separation of crude CB[n] mixture on the laboratory (i.e., up to 1 kg) to industrial (i.e., tonnes) scale must be developed. Alternatively, synthetic procedures that selectively target a speciﬁc CB[n] (e.g., CB[8] or CB[10]) would be tremendously valuable. Second, the development of high-yielding methods of functionalizing pre-formed CB[n] – particularly for CB[7], CB[8], and CB[10] – would have a dramatic impact on the ﬁeld. Third, a deﬁciency of the CB[n] family relative to the cyclodextrins is that they are inherently achiral, which

4.8 Experimental: Selected Procedures

renders chiral recognition inside CB[n] challenging. Our group and Inoues group have recently shown that chiral recognition inside CB[n] can be achieved by assembled enantioselection, where a chiral binary complex can be transformed into a diastereomerically enriched ternary complex by interaction with a chiral but racemic guest [18,43]. Fourth, to date a small number of studies have targeted the development of CB[n] for electronic applications [44]. Fifth, for applications involving humans (e.g., drug delivery, foodstuff additives, perfume additives, etc.) it will be necessary to demonstrate that CB[n] compounds are not toxic and that they are properly excreted from the body. Given the high chemical stability of CB[n] it seems unlikely that CB[n] will be metabolized in vivo. Lastly, it seems likely that CB[n] will be equal or superior to the cyclodextrins in the numerous application areas that have been demonstrated, and much future work will be directed accordingly.

4.8 Experimental: Selected Procedures

This experimental section describes several key procedures encountered during the synthesis of members of the CB [n] family. 4.8.1 Synthesis of Glycolurils

An excellent review by Petersen describes the preparation of numerous glycolurils [45]. In general, the formation of glycolurils occurs by the condensation of urea and its derivatives with 1,2-diketones under acidic conditions. Two speciﬁc examples are presented below which use either aqueous or anhydrous acidic conditions. Synthesis of Glycoluril 1a A modiﬁcation of the literature procedure was used [33]. To a solution of urea (600 g, 10 mol) in water (1 L) was added a 40% aq. solution of glyoxal (500 g, 3.45 mol) and conc. HCl (86 mL). The resulting solution was heated at 85–90 C until a heavy precipitate was formed. The reaction mixture was allowed to cool to room temperature and ﬁltered. The ﬁlter cake was washed with copious amounts of water (2 L) followed by acetone to remove residual water. The resulting white solid (397.5 g, 81%) was dried under high vacuum. 1 H NMR (DMSO, 400 MHz): 7.11 (s, 4H), 5.20 (s, 2H). Synthesis of Glycoluril 1b A modiﬁcation of the literature procedures was used [46]. A mixture of benzil (21.0 g, 0.10 mol) and urea (12.1 g, 0.20 mol) in benzene (380 mL) was treated with TFA (20 mL) which resulted in a homogeneous mixture. The reaction mixture was then heated overnight at reﬂux with azeotropic removal of water. After cooling to room temperature, the white precipitate was isolated by ﬁltration, washed with EtOH, and dried under high vacuum yielding 1b as a white solid (22.8 g, 78%). 1 H NMR (DMSO, 400 MHz): 7.70 (s, 4H), 7.05-6.95 (m, 10H).

j137

138

j 4 Cucurbit[n]urils 4.8.2 Synthesis and Separation of Cucurbit[n]urils

Although the synthesis and separation procedures given below may appear straightforward, there are numerous details which affect the outcome of the CB[n]-forming reaction (e.g., rate of stirring, physical state and degree of physical mixing of the starting materials, homogeneity of the solution, etc.) and the efﬁciency of the separation procedure (e.g., HCl content in the crude solid). Therefore, each of the major laboratories involved in CB[n] chemistry has developed its own modiﬁcations of the basic procedures published by Day and Kim [12,15]. We present here the procedures used by our group at the University of Maryland. Synthesis of CB[n] and i-CB[n] Powdered glycoluril (795 g, 5.59 mol) and powdered paraformaldehyde (354 g, 11.2 mol) were mixed thoroughly. An ice-cold concentrated HCl solution (100 mL) was added gradually while stirring with a large glass rod. The reactants were transformed into a brick-like material, stirring was no longer possible, and the reaction became highly exothermic. An additional portion of icecold concentrated HCl (1030 mL) was added gradually. The heterogenous mixture was heated to 80 C over 2.5 h and maintained at that temperature for an additional 2.5 h, at which point the solid had dissolved. The homogeneous red solution was heated to 100 C for 14 h. After cooling to room temperature, the puriﬁcation process was begun. Monitoring of the Purification Process It is possible to monitor the puriﬁcation process by 1 H NMR using 20% DCl as the solvent. In this solvent CB[5]–CB[8], CB[10] sCB[5], and i-CB[6] show diagnostic resonances for the high-ﬁeld

Figure 4.3 1H NMR spectrum (400 MHz, 20 % DCl, RT) of the crude CB[n] reaction mixture.

4.8 Experimental: Selected Procedures

diastereotopic CH2-group (4.5–4.1 ppm). Figure 4.3 shows a representative 1 H NMR spectrum of the crude CB[n]-forming reaction mixture. Purification of CB[5], CB[6], CB[7], CB[8], CB[10] sCB[5], and i-CB[6] The reaction mixture, which contained a large amount of solid, was evaporated to a minimum volume (600 mL). This slurry was poured into water (2.5 L). The solid was collected by ﬁltration (Crop 1 contains CB[6], CB[7], CB[8], some i-CB[6], and some CB[10] s CB[5]). The ﬁltrate was evaporated to about 600 mL and then slowly poured into a mixture of MeOH (3 L) and water (200 mL) with vigorous stirring. After stirring overnight, the precipitate was isolated by ﬁltration (Crop 2 contains mainly CB[7] and CB[5] and smaller amounts of CB[6]). Subsequent Purification The separation of each component (CB[5], CB[6], CB[7], CB[8], CB[10] sCB[5], and i-CB[6]) from Crop 1 and Crop 2 takes advantage of their differential solubility in HCl solutions of different concentration. For example, CB[5] and CB[7] have moderate solubility in water but other CB[n] compounds are nearly insoluble in pure water. The crude solids are stirred in a beaker with large volumes of water (2 L) repeatedly to isolate a solution consisting of mainly CB[5] and CB[7]. After concentration to 200 mL, the solution is poured into MeOH (1.5 L) and the solid isolated by ﬁltration. Separation of CB[5] from CB[7] is based on its moderate solubility (about 33 mg mL1) in 50% aq. MeOH solution (v/v). The solubility of CB[7] is less than 4 mg mL1 in this solution. Accordingly, the solid mixture of CB[5] and CB[7] is repeatedly stirred in a large beaker with 50% aq. MeOH to dissolve the CB[5] and leave the CB[7] as a solid. The solution of CB[5] is concentrated (not to dryness), which yields crystalline CB[5] (8% overall yield). Diffusion of acetone into an aq. soln. of CB[7] is used to obtain CB[7] as a crystalline solid (25% overall yield). The solid material from which CB[5] and CB[7] has been removed is subsequently processed to yield CB[6], i-CB[6], CB[8], and CB[10] sCB[5]. In 3 M HCl, CB[6], i-CB [6], and CB[10] sCB[5] have appreciable solubility, whereas CB[8] is substantially less soluble. By stirring the crude mixture of CB[6], CB[8], CB[10] sCB[5], and i-CB[6], with 3 M HCl in a large beaker it is possible to isolate CB[8] as an insoluble solid. CB [8] can then be recrystallized from warm conc. HCl to yield a crystalline solid (10% overall yield). The solution of CB[6], i-CB[6], and CB[10] sCB[5] in 3 M HCl is concentrated by rotary evaporation resulting in precipitation of CB[6]. The CB[n] compounds remaining in solution are precipitated by pouring the aq. soln. into MeOH. The solid obtained, which still contains CB[6], i-CB[6], and CB[10] sCB[5], can be separated by repeated fractional crystallization from HCl solutions of different concentrations. For example, CB[10] sCB[5] can be obtained by crystallization of the mixture from conc. HCl (2% overall yield). In samples that contain a large amount of i-CB[6], the CB[6] impurity can be removed by dissolving the mixture in 20% HCl, adding the calculated amount of hexanediamine, and diluting 10-fold with water. i-CB[6] is obtained as an insoluble precipitate (2% overall yield). CB[6] is obtained from multiple fractions and can be recrystallized from concentrated HCl (50% overall yield). The synthesis and puriﬁcation procedure described above takes approximately 6 weeks.

j139

140

j 4 Cucurbit[n]urils 4.8.3 Synthesis of Nor-Seco-Cucurbit[10]uril [14]

A mixture of glycoluril (1.42 g, 9.99 mmol), paraformaldehyde (0.50 g, 16.69 mmol), and conc. HCl (4 mL) was heated at 50 C for 3 days. The resulting precipitate was separated by centrifugation to yield a crude solid (300 mg). This was dried under high vacuum for 10 min, washed with HCl:H2O (1:1, v/v), washed with MeOH to remove HCl, and dried under high vacuum overnight to yield ns-CB[10] (238 mg, 0.145 mmol, 15%) as a white solid. 1 H NMR (400 MHz, 20% DCl): 5.47 (m, 28H), 5.40 (s, 4H), 5.33 (d, J ¼ 4.0, 4H), 4.60 (s, 4H), 4.20 (m, 16H).

Acknowledgment

We thank the past and current members of the Isaacs group for their contributions to the CB[n] project. We also thank Professors Brian Wagner, Kimoon Kim, Angel Kaifer, and Yoshihisa Inoue as well as Dr. Yiu-Fai Lam (NMR), Dr. Peter Y. Zavalij, and James C. Fettinger (X-ray) for their collaboration in various aspects of the research described in this chapter. Finally, we thank the National Science Foundation (CHE-0615049), National Institutes of Health (GM61854), the Petroleum Research Fund administered by the American Chemical Society (PRF-45228-AC4), and the University of Maryland for ﬁnancial support of the research described in this chapter.

Abbreviations

aq. CB[n] Conc. d DMF DMSO equiv EYPC HPTS hv i-CB[n] J Ka LUVs m MMFF MV n.d. NMR

aqueous cucurbit[n]uril concentrated doublet dimethylformamide dimethylsulfoxide equivalents egg yolk L-a-phosphatidyl-choline 8-hydroxypyrene-1,3,6-trisulfonate light inverted cucurbit[n]uril coupling constant (Hz) equilibrium association constant large unilamellar vesicles multiplet Merck molecular force ﬁeld methyl viologen not detected nuclear magnetic resonance

References

ns-CB[10] PTSA s TFA v/v

bis-nor-seco-cucurbit[10]uril p-toluenesulfonic acid singlet triﬂuoroacetic acid volume to volume ratio

References 1 Behrend, R., Meyer, E. and Rusche, F. (1905) Liebigs Ann. Chem., 339, 1–37. 2 Freeman, W.A., Mock, W.L. and Shih, N.-Y. (1981) J. Am. Chem. Soc., 103, 7367–7368. 3 Flinn, A., Hough, G.C., Stoddart, J.F. and Williams, D.J. (1992) Angew. Chem., 104, 1550–1552; Angew. Chem., Int. Ed. Engl., 31, 1475–1477. 4 Wu, A., Chakraborty, A., Witt, D., Lagona, J., Damkaci, F., Ofori, M.A., Chiles, J.K., Fettinger, J.C. and Isaacs, L. (2002) J. Org. Chem., 67, 5817–5830. 5 Chakraborty, A., Wu, A., Witt, D., Lagona, J., Fettinger, J.C. and Isaacs, L. (2002) J. Am. Chem. Soc., 124, 8297–8306. 6 Witt, D., Lagona, J., Damkaci, F., Fettinger, J.C. and Isaacs, L. (2000) Org. Lett., 2, 755–758. 7 Mock, W.L. and Shih, N.-Y. (1983) J. Org. Chem., 48, 3618–3619; J. Am. Chem. Soc., 110, 4706–4710. 8 Mock, W.L. and Shih, N.-Y. (1986) J. Org. Chem., 51, 4440–4446. 9 Mock, W.L. and Shih, N.-Y. (1989) J. Am. Chem. Soc., 111, 2697–2699. 10 Mock, W.L., Irra, T.A., Wepsiec, J.P. and Manimaran, T.L. (1983) J. Org. Chem., 48, 3619–3620; Mock, W.L., Irra, T.A., Wepsiec, J.P. and Adhya, M. (1989) J. Org. Chem., 54, 5302–5308. 11 Mock, W.L. and Pierpont, J. (1990) J.Chem.Soc.Chem.Commun., 1509–1511. 12 Day, A., Arnold, A.P., Blanch, R.J. and Snushall, B. (2001) J. Org. Chem., 66, 8094–8100. 13 Isaacs, L., Park, S.-K., Liu, S., Ko, Y.H., Selvapalam, N., Kim, Y., Kim, H., Zavalij, P.Y., Kim, G.-H., Lee, H.-S. and

14

15

16

17

18 19

20 21

22

23

24

25

Kim, K. (2005) J. Am. Chem. Soc., 127, 18000–18001. Huang, W.-H., Liu, S., Zavalij, P.Y. and Isaacs, L. (2006) J. Am. Chem. Soc., 128, 14744–14745. Kim, J., Jung, I.-S., Kim, S.-Y., Lee, E., Kang, J.-K., Sakamoto, S., Yamaguchi, K. and Kim, K. (2000) J. Am. Chem. Soc., 122, 540–541. Lee, J.W., Samal, S., Selvapalam, N., Kim, H.-J. and Kim, K. (2003) Acc. Chem. Res., 36, 621–630. Day, A.I., Blanch, R.J., Arnold, A.P., Lorenzo, S., Lewis, G.R. and Dance, I. (2002) Angew. Chem., 114, 285–287; Angew. Chem., Int. Ed., 41, 275–277. Liu, S., Zavalij, P.Y. and Isaacs, L. (2005) J. Am. Chem. Soc., 127, 16798–16799. Liu, S., Ruspic, C., Mukhopadhyay, P., Chakrabarti, S., Zavalij, P.Y. and Isaacs, L. (2005) J. Am. Chem. Soc., 127, 15959–15967. Mukhopadhyay, P., Wu, A. and Isaacs, L. (2004) J. Org. Chem., 69, 6157–6164. Mukhopadhyay, P., Zavalij, P.Y. and Isaacs, L. (2006) J. Am. Chem. Soc., 128, 14093–14102. Jun, S.I., Lee, J.W., Sakamoto, S., Yamaguchi, K. and Kim, K. (2000) Tetrahedron Lett., 41, 471–475. Moon, K., Grindstaff, J., Sobransingh, D. and Kaifer, A.E. (2004) Angew. Chem., 116, 5612–5615; Angew. Chem., Int. Ed., 43, 5496–5499. Ong, W. and Kaifer, A.E. (2003) Angew. Chem., 115, 2214–2217; Angew. Chem., Int. Ed., 42, 2164–2167. Lim, Y.-B., Kim, T., Lee, J.W., Kim, S.-M., Kim, H.-J., Kim, K. and Park, J.-S. (2002)

j141

142

j 4 Cucurbit[n]urils

26

27

28

29

30

31

32

33

Bioconjugate Chem., 13, 1181–1185; Lee, J.W., Ko, Y.H., Park, S.-H., Yamaguchi, K. and Kim, K. (2001) Angew. Chem., 113, 768–771; Angew. Chem., Int. Ed., 40, 746– 749. Jeon, W.S., Ziganshina, A.Y., Lee, J.W., Ko, Y.H., Kang, J.-K., Lee, C. and Kim, K. (2003) Angew. Chem., 115, 4231–4234; Angew. Chem., Int. Ed., 42, 4097–4100; Ko, Y.H., Kim, K., Kang, J.-K., Chun, H., Lee, J.W., Sakamoto, S., Yamaguchi, K., Fettinger, J.C. and Kim, K. (2004) J. Am. Chem. Soc., 126, 1932–1933. Jeon, W.S., Kim, E., Ko, Y.H., Hwang, I., Lee, J.W., Kim, S.-Y., Kim, H.-J. and Kim, K. (2005) Angew. Chem., 117, 89– 93; Angew. Chem., Int. Ed., 44, 87–91. Marquez, C. and Nau, W.M. (2001) Angew. Chem., 113, 3248–3253; Angew. Chem., Int. Ed., 40, 3155–3160. Jeon, Y.J., Kim, S.-Y., Ko, Y.H., Sakamoto, S., Yamaguchi, K. and Kim, K. (2005) Org. Biomol. Chem., 3, 2122–2125. Wheate, N.J., Day, A.I., Blanch, R.J., Arnold, A.P., Cullinane, C. and Collins, J. G. (2004) Chem. Commun., 1424–1425; Wheate, N.J., Buck, D.P., Day, A.I. and Collins,J.G.(2006) Dalton Trans.,451–458. Rowan, A.E., Elemans, J.A.A.W. and Nolte, R.J.M. (1999) Acc. Chem. Res., 32, 995–1006; Rebek, J. Jr. (1999) Acc. Chem. Res., 32, 278–286. Eliel, E.L. and Wilen, S.H. (1994) Stereochemistry of Organic Compounds, J. Wiley and Sons, New York. Buschmann, H.-J., Fink, H. and Schollmeyer, E. (1997) Preparation of cucurbituril” Ger. Offen. [Germany]: DE 19603377 [Chem. Abstr. 1997, 127, 205599].

34 Lagona, J., Fettinger, J.C. and Isaacs, L. (2003) Org. Lett., 5, 3745–3747. 35 Lagona, J., Fettinger, J.C. and Isaacs, L. (2005) J. Org. Chem., 70, 10381–10392. 36 Lagona, J., Wagner, B.D. and Isaacs, L. (2006) J. Org. Chem., 71, 1181–1190. 37 Day, A.I., Arnold, A.P. and Blanch, R.J. (2003) Molecules, 8, 74–84. 38 Zhao, Y., Xue, S., Zhu, Q., Tao, Z., Zhang, J., Wei, Z., Long, L., Hu, M., Xiao, H. and Day, A. I. (2004) Chin. Science Bull., 49, 1111–1116. 39 Jon, S.Y., Selvapalam, N., Oh, D.H., Kang, J.-K., Kim, S.-Y., Jeon, Y.J., Lee, J. W. and Kim, K. (2003) J. Am. Chem. Soc., 125, 10186–10187. 40 Lee, H.-K., Park, K.M., Jeon, Y.J., Kim, D., Oh, D.H., Kim, H.S., Park, C.K. and Kim, K. (2005) J. Am. Chem. Soc., 127, 5006–5007. 41 Jeon, Y.J., Kim, H., Jon, S., Selvapalam, N., Oh, D.H., Seo, I., Park, C.-S., Jung, S.R., Koh, D.-S. and Kim, K. (2004) J. Am. Chem. Soc., 126, 15944–15945. 42 Rudkevich, D.M. (2004) Angew. Chem., 116, 568–581; Angew. Chem., Int. Ed., 43, 558–571. 43 Rekharsky, M.V., Yamamura, H., Inoue, C., Kawai, M., Osaka, I., Arakawa, R., Shiba, K., Sato, A., Ko, Y.H., Selvapalam, N., Kim, K. and Inoue, Y. (2006) J. Am. Chem. Soc., 128, 14871–14880. 44 Ling, Y. and Kaifer, A.E. (2006) Chem. Mater., 18, 5944–5949. 45 Petersen, H. (1973) Synthesis, 243–292. 46 Butler, A.R. and Leitch, E. (1980) J. Chem. Soc. Perkin Trans. 2, 103–105.

j143

5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes Ganna Podoprygorina and Volker Bo¨hmer

5.1 Introduction

The preparation of topologically interesting molecules and supramolecular structures such as catenanes, knots [1], or Möbius strips [2] is primarily an intellectual challenge, although more and more applications [3] have been proposed during the last decade(s). The synthesis of structures consisting of interlocked rings usually requires the formation of cyclic molecules from linear precursors while they are penetrating already existing rings. Similarly to the macroscopic example (the fabrication of a chain [Latin: catena]), an efﬁcient synthesis of catenanes1) requires a perfect prearrangement of the respective building blocks. On the molecular level, such a preorganization can be achieved by an appropriate covalent linkage which has to be cleaved after the intended ring-closure reaction(s). However, the desired mutual arrangement of the precursors can also be obtained via reversible links, such as coordination to metal cations, hydrogen bonding, or interactions of the donoracceptor type. This has the advantage that wrong connections can be corrected during the assembly process. The rapid progress in the synthesis of topologically more and more demanding molecules and complex structures in reasonable yields, for which Borromean rings [4] represent an impressive example, is therefore strongly connected with the development of supramolecular chemistry in general [5]. The ﬁrst rational synthesis of a [2]catenane was reported by Schill and Lüttringhaus [6], who successfully used a covalent attachment to a template. This covalent approach was recently applied by Godt and coworkers for the preparation of a series of [2]catenanes consisting of really large and thus mutually mobile chain rings (non-rusty catenanes) [7]. In this case a prefabricated macrocycle and a precursor for the second ring were linked via a carbonate group. The oxidative coupling of the alkyne residues formed the second cycle, which has to be threaded through the ﬁrst one (Scheme 5.1). Hydrolysis of the carbonate linkage disconnected the two rings, producing a [2]catenane. 1) Similar considerations hold for the topologically trivial rotaxanes.

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

144

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes X

O

X X

O O

O

X

O EtO

O

O

OEt

O X

EtO

HO

OH

OE t

X

O

X X

O

X = (CH2)11, (CH2) 23

O

Scheme 5.1 Covalent prearrangement in the synthesis of [2]catenanes consisting of large, mobile rings. The structure of the ﬁrst ring should not allow an exo orientation of the carbonate link.

Various catenanes [8], rotaxanes [8], and even trefoil knots [9] were synthesized in high yields using the coordination of suitably functionalized precursors as ligands around metal ions. The tetrahedral organization of phenanthroline derivatives by copper(I) was the ﬁrst [10] and still remains [11] one of the most frequently used motifs. The synthesis of a bis[2]catenane [12] in which a large ring formed by metathesis reaction is threaded through two rings of an eight-shaped bis-macrocycle (Scheme 5.2) is a recent example. Metal cations with an octahedral coordination sphere have also been utilized as templates for the directed synthesis of various interlocked structures [13], including Borromean rings [4,14]. Terpyridyl or 2,6-diiminopyridine derivatives (Scheme 5.3) self-organize into complexes in the presence of metal cations, providing an arrangement of the attached functional groups X suitable for their selective ring-closure to [2]catenanes. The same catenanes consisting of diiminopyridine rings can be also prepared from the corresponding diamines and 2,6-pyridine dialdehyde [13b]. In this case, reversible Schiff base formation favors the thermodynamically stable species and leads ﬁnally to those products which complete the octahedral coordination sphere of the templating cation. 2+ O

O O

N

O

N

N

O

2+

Cu N

N

O O

O

O O

O

O

O N N Cu N N

N

O

O

O

O

O

O

O

O

N

O

N

N

N

N

N

O

Scheme 5.2 Synthesis of a bis[2]catenane templated by two copper (I) ions. The starting complex is the only species which enables a complete interaction of all phenanthroline sites with Cu(I).

O N N Cu N N

Cu O

O

O

O O

O

O O

5.1 Introduction 2+ X

N

O N

N Y

N

M

N

2+

X

N

N

N

X

NH2

2+

M M

N N

X

j145

Y

Y

N

H +

NH2

N

N H O

Scheme 5.3 Synthesis of [2]catenanes by coordination of 2,6disubstituted pyridine ligands to cations M2þ ¼ Mn2þ, Fe2þ, Co2þ, Ni2þ, Cu2þ, Zn2þ, Cd2þ, Hg2þ. Diimines are kept by the template in an orthogonal arrangement which leads to a selective ring closure. X symbolizes various reactive groups used for the covalent connection of the preorganized building blocks by an appropriate spacer Y.

Anions have also been used as templates in the synthesis of catenanes and rotaxanes, as recently described by Beer and coworkers [15]. They used the selfassembly of diamides, known as anion receptor molecules, around a chloride anion to produce, for example, [2]- and [3]catenanes by oleﬁn metathesis [16] (Scheme 5.4). In this special example, the template effect was increased by hydrogen bonding between the N-methyl group to the crown ether chain and by p-p stacking interactions. However, preorganization by reversible interactions via hydrogen bonds does not necessarily need a further template such as an anion. Direct intermolecular hydrogen bonds between derivatives of isophthalic acid and 2,6-pyridine diacid were used to synthesize a series of new catenanes [17,18], rotaxanes [18c,19], and trefoil knots [20]. The interaction of dibenzylammonium cations and dibenzo[24]crown[8] [21] led to the preparation of rotaxanes by the attachment of stoppers to the thread ends (Scheme 5.5) [22], or similarly by clipping of the crown ether chain wrapped around the ammonium dumbbell [23]. Recently, several novel catenanes were synthesized via the macrocyclic connection of three ammonium cations threaded through three crown ether rings attached to a triptycene core [24]. Numerous interlocked structures [25], including a [5]catenane (olympiadane) [26], were built using the self-organization of crown ether derivatives containing O

O

O O

H H H

O

+ N

Cl O

O

O

HN

NH

NH O

O

O

O

O

-

HN

H H H

O

O

O

O

O

O

O

O

O

Scheme 5.4 Synthesis of a [2]catenane by self-organization of amides around a chloride anion as the template. The pseudorotaxane formed between the linear and the cyclic amides is held together additionally by various hydrogen bonds and p-p interactions.

+ N

Cl NH O

O

O

HN

NH

O

O

O

O

O

O

O

O

HN

O

N3

CH 3 CN

PF 6

-

N3

N3

O

O + N

O

O

O

H H

O

Scheme 5.5 A pseudorotaxane is formed by hydrogen bonding between benzylic protons and the -NH2- protons of the ammonium cation with the oxygens of a crown ether. The introduction of bulky residues to appropriately functionalized thread ends produces a rotaxane, while catenanes are accessible by a cyclic connection of the ends.

O

O

O

O

O

O

O

O

+ N H2

-

PF 6

O

O

N3

O

O

O

O

O

O

O

O

N N

N

O

O

+ N

O

O

O

H H

O

PF 6

O

O

N

N

N

O

O

O

O

146

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

5.1 Introduction O O O

O

O O

+N

N

O

O

O O

O

Br N+

N 2PF6-

Br

O

O

+N

O

O

O

O

+N

O

N+

N+ 2PF6- + 2Br-

O

Scheme 5.6 The general principle for the synthesis of catenanes (or rotaxanes) based on cyclobis(paraquat-p-phenylene) and crown ethers containing hydroquinone units. The dipyridyl units are alternatingly stacked with two hydroquinones because of p-p interactions which direct the ring-closure toward the formation of a catenane.

hydroquinone (or similar p-electron-donating) units and bis(paraquat-p-phenylene) (Scheme 5.6) or its analogs. In this case, p-p interactions between a hydroquinone donor and a bipyridyl acceptor unit, enhanced by hydrogen bonding between the benzylic protons and the crown ether oxygens, are responsible for the preorganization. As a general principle, effective syntheses of interlocked molecules are based on the supramolecular preorganization of the starting cyclic and linear components. These (more or less) stable intermediate assemblies are referred to as pseudorotaxanes [27]. Depending upon the functionalization, the same pseudorotaxane can be a precursor for catenanes (by ring-closure) and for rotaxanes (by appropriate stoppering). Pseudorotaxanes or rotaxanes can be also obtained via clipping, the cyclization of a chain assembled around a linear or dumbbell shaped molecule correspondingly. These possibilities are schematically summarized in Scheme 5.7.

Ring-closure

Threading

Pseudorotaxane Catenane Stoppering

Slipping

Rotaxane

Clipping

Scheme 5.7 Pathways leading to the formation of rotaxanes or catenanes.

j147

148

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes Hydrogen-bonded dimers of tetra-urea calix[4]arenes offer a unique, unprecedented preorganization [28]. Before we describe how this can be used for the rational construction of molecules with novel topologies, it is reasonable to discuss the properties of such dimeric capsules and the conditions under which they exist.

5.2 Basics of Tetra-urea Calix[4]arenes 5.2.1 Synthesis

Starting from tert-butylcalix[4]arene [29], the preparation of a typical tetra-urea derivative usually requires four steps (Scheme 5.8): a. Exhaustive O-alkylation to ﬁx the cone conformation; b. Substitution of t-butyl groups by ipso-nitration; c. Reduction of the nitro groups; d. Acylation by an appropriate isocyanate. The isocyanate can be replaced by the corresponding activated urethane (as a milder reagent), and this activation can also involve the p-amino functions of the calix[4]arene (step e) and their subsequent reaction with a suitable amine (step f). Multigram quantities of simple tetra-urea calix[4]arenes are easily available in this way. Selective substitution by different urea groups is easily accomplished using protection-deprotection strategies. Mono-, 1,2-di-, and tri-Boc derivatives are available in gram quantities [30], which allows the preparation of AAAB- and AABB-type

d Y

Y

O

O

NO2

4

4

NH2

NH

4

a

O

f

e

O C Y =H

Y

O

c

b t-Bu

Y

Y

O

NH

4 O C

O

NH R

Y = Alk yl NO2

Scheme 5.8 Synthesis of tetra-urea calix[4]arenes. (a) Alkylating reagents, NaH, DMF, rt; (b) HNO3/HOAc, CH2Cl2, rt; (c) H2, Raney-Ni, toluene, rt; (d) R-NCO, rt; (e) p-nitrophenyl chloroformate; (f) R-NH2, rt.

4

5.2 Basics of Tetra-urea Calix[4]arenes Y

Y O

O

NH

NH2

Y

Y

Y

Y

O

O

O

O

NH

NH 2 C O NH

j149

a 2

Y O

HN HN

R1

NH2

R1

R2

4 O

O

HN O

Y

Y

Y

Y

b

HN 2 C O C O HN

O

NH2 HN C O C O O

O

NH2

HN O

O

O

O

HN HN HN C O C O C O C O HN HN O R1

R1

O

HN HN R2

O

Y

Y

Y

Y

Y

Y

Y

Y O

O

O

HN HN HN C O C O C O C O HN HN HN R2

R1

R1

Scheme 5.9 Synthesis of tetra-urea calix[4]arenes bearing two different urea residues in the sequence ABAB (a) and AABB (b), using the protection of amino functions by trityl- or Boc-groups. (Y ¼ C5H11).

tetra-ureas.2) The ABAB-type is now also readily accessible, using the protection of distal (opposite) amino groups by alkylation with tritylchloride [31]. Examples of these syntheses are shown in Scheme 5.9. 5.2.2 Dimeric Capsules

As ﬁrst described by Shimizu and Rebek Jr. [32], tetra-urea calix[4]arenes form dimeric capsules in aprotic, apolar solvents such as chloroform, cyclohexane, or benzene, while they exist as single molecules in polar solvents such as acetone, DMSO, pyridine, or ethanol [33]. This was ﬁrst deduced from the 1 H NMR spectra (compare Figure 5.1a,b) which show for the dimers a strong downﬁeld shift for one of the two NH-signals, indicating strong hydrogen bonds, and a splitting of the singlet of the aromatic protons of the calix[4]arene skeleton into two m-coupled doublets, indicating a change in symmetry from C4v to C4. The structure of the capsule, proposed initially on the basis of these NMR-data, was completely conﬁrmed by the ﬁrst single crystal X-ray structure (Figure 5.2) [34]. The two calix[4]arenes are connected via their wide rims by a seam of intermolecular hydrogen bonds alternatingly involving the urea residues of both calixarenes. The directionality of this hydrogen bonded belt is the reason for the inequivalence of the two aromatic protons of a given phenolic unit mentioned above. The internal volume of the capsule (190–200 Å3) cannot be empty, and the inclusion of a suitable guest (often the solvent) is a necessary condition for the dimerization. Various organic molecules (for example, (1R)-(þ)-nopinone, (1R)-(þ)-camphor, [32b,c]) have been included in the cavity [35], and the upﬁeld shifts of their 1 H NMR signals provide an independent proof of the formation of dimeric capsules, 2) A and B just characterize different phenolic units, where the difference may be due to the ether residues

Y or to the urea residues R or both.

150

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes Ar H

(a) NH

C5H11 O

*

(b) NH

NH

Ar H

Ar H

H

H NH

4

O C NH

(c)

*

NH

δ/ppm 9.0

8.0

1

7.0

6.0

Figure 5.1 Sections of 1H NMR spectra (400 MHz, 25 C) of tetra-urea 1: (a) in DMSO-d6 (monomers); (b) in CDCl3 (dimeric capsules 1CDCl31); (c) in CDCl3 in the presence of Et4NþBF4 (dimeric capsules [1Et4Nþ1]BF4—). The solvent signal is marked by an asterisk.

although these guests are usually not found by X-ray diffraction because of their disorder. Organic cations are especially good guests, since their inclusion is favored by cation-p interactions with the aromatic walls. This enables the encapsulation to be studied by mass spectrometry also [36]. If cations such as tetraethylammonium or cobaltocenium are included [37], the directionality of the hydrogen-bonded belt

Figure 5.2 Molecular structure of a tetra-urea dimer 22 determined by X-ray analysis [34]. (a) Side view; (b) top view. Alkyl residues (in b), hydrogen atoms, and solvent molecules are omitted for clarity.

5.2 Basics of Tetra-urea Calix[4]arenes

j151

changes rapidly on the NMR-time scale. The typical m-coupled doublets for the aromatic protons, usually observed even at elevated temperatures, collapse into a singlet (Figure 5.1c). Neutral molecules of similar size (tetraethylsilane, ferrocene) are not accepted as guest, which demonstrates the importance of cation-p interactions for the stability of the whole assembly. Up to now, a dimer with tetraethylammonium is the only example where the guest was located in the X-ray structure [38]. 5.2.3 Heterodimers

A dimeric capsule can also be formed by two different tetra-urea molecules, and in aprotic solvents a mixture of two tetra-urea derivatives usually contains not only the two homodimers but also the heterodimer in a more or less statistical ratio. This formation of heterodimers is an additional proof of the dimerization [39], which is also valid when other indications (e.g. m-coupled doublets for aryl protons) fail. As an example, sections of the spectra of tetra-tolylurea (1), tetra-hexylurea (3), and of a mixture of the two are shown in Figure 5.3a–c [37a]. However, the statistical formation of heterodimers is not the rule. Examples are known where no heterodimers are formed, e.g. between 1 and tetra-ureas derived from a rigid bis-crown-3 [40] (see Section 7). Further examples will be shown in the next sections. In such cases, the solution of the two tetra-urea calix[4]arenes contains exclusively the two homodimers.

(a)

C10H21

C5H11

O

O

(b)

NH O C

NH

4

O C NH

NH

(c) 1 4.5

4.0

3.5

3.0

Figure 5.3 Sections of 1H NMR spectra (CDCl3, 400 MHz, 25 C) of tetra-urea dimers with tetraethylammonium as guest: (a) [1Et4Nþ1]BF4; (b) [3Et4Nþ3]BF4; (c) 1:1 mixture of [1Et4Nþ1]BF4, and [3Et4Nþ3]BF4. In addition to the signals for the two homodimers (a, b), the mixture (c) contains a double set of signals for the two parts of the heterodimer.The signals of free Et4Nþ are marked by a circle.

δ/ppm

3

4

152

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

Y

C5H11

O

O

(a) NH

NH

4

O C

4

O C NH

NH O S O

(b)

(c)

1 4

δ /ppm 10.0

9.0

8.0

7.0

6.0

4.9

1

Figure 5.4 Representative sections of the H NMR-Spectra (CDCl3, 400 MHz, 25 C) of (a) tetra-tolylurea (1) (Y ¼ C5H11), (b) tetra-tosylurea (4), and (c) a 1:1 mixture of 1 and 4.

On the other hand, there are examples of exclusive heterodimerization. Both tetraaryl- and tetra-tosylureas readily form homodimers when they are dissolved alone in an apolar solvent. However, heterodimers are the only detectable species in a solution containing both compounds in a 1:1 ratio. An example of this selective heterodimerization [41] is demonstrated by the NMR-spectra in Figure 5.4. Only when one of these tetra-ureas is present in excess can signals for its homodimer be seen by NMR. This early observation by Rebek and Castellano [42] is especially important for the template syntheses described below. Other examples of exclusive formation of heterodimers will be discussed later. 5.2.4 Symmetry Properties 1

H NMR spectra still represent the main source of information for more sophisticated dimeric assemblies also. To understand these, it is useful to discuss some symmetry properties of hetero- and homodimers ﬁrst, which are the basis for their characterization by 1 H NMR. Figure 5.5 summarizes the main examples. A homodimer of a tetra-urea calix[4]arene consisting of identical phenolic units A is composed of two enantiomers with C4-symmetry, which results in overall S8symmetry. Consequently, a heterodimer with a second calixarene consisting of four units B must be chiral, but this chirality is due only to the directionality of the hydrogen-bonded belt or (in other words) to the orientation of the carbonyl groups [42,43]. Rotation around the (four) aryl-NH bonds leads to the opposite enantiomer (conformational chirality). A single calixarene consisting of two different phenolic units A and B in alternating order ABAB is (time averaged) C2v-symmetric. Consequently, its homodimer is chiral (D2-symmetry) without the directionality of the hydrogen-bonded belt, just by

5.3 Preorganization in Dimers of Tetra-urea Calix[4]arenes

Heterodimer

Homodimer σ σ σ

A

A

σ A

A σ

σ A σ

A

A A σ

A

A A

A A

σ

σ

A

A

σ

A

σ

B

σ A

A

D4d

B

B

A σ B

S8

B A σ

A

B

A

A B

σ

B A

C4v

C4

Homodimers ABAB-Type A

A

B

B

B

B B

ABBB-Type A

B

A

B B

A

A

B B

B

A A

B B

B

C2

A B

C1

B

A

B B

B

B B

C1

A

A A

B

A

A B

A B

C2

A

B A B

C1

B

B

A

C2

A

B

B A

B

B

A

A

B

B

C2

B

B

A B

B

B

A

AABB-Type

B

C2

B

B

B

A

D2 A

B

A

B

A

B

B

A A

A B

C1

Figure 5.5 Survey of the symmetry properties for selected types of homo- and heterodimers of tetra-urea calix[4]arenes, represented by squares with the phenolic units (A, B) on the corners. Symmetry elements and symmetry classes (with and without directionality of the hydrogen bonds, shown by arrows) are indicated.

the mutual arrangement of the two non-chiral calix[4]arenes (supramolecular chirality) [44]. The directionality of the hydrogen bonds does not create new stereoisomers; it just reduces the symmetry from D2 to C2. For tetra-urea calix[4]arenes of the AABB- and ABBB-types, two regioisomeric dimers are possible, which both have C2- or C1-symmetry, respectively.

5.3 Preorganization in Dimers of Tetra-urea Calix[4]arenes 5.3.1 General Considerations

The inclusion of guests is not the only interesting aspect of the dimers formed by tetra-urea calixarenes. Numerous other self-assembled capsules are known, larger in volume, different in shape, and able to include more than one guest molecule, and various aspects of guest inclusion have been extensively studied [45].

j153

154

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

Figure 5.6 Hydrogen-bonded dimer of a tetra-tolylurea calix[4] arene. Stick models based on molecular modeling (a) and X-ray structures (b), illustrating possible intra- and intermolecular connections (a- and b-connections).

The dimerization also brings the eight residues attached to the urea groups into a unique mutual position. This means that the residues are preorganized for controlled reactions between them, for instance, for their intramolecular connection. As illustrated in Figure 5.6 for a tetra-tolylurea (as an example), the two calix[4]arenes forming the dimer are turned by 45 around their common fourfold axis. The tolyl residues attached to the urea groups point in opposite directions. Based on MDsimulations, the distance between adjacent tolyl methyl groups in the same calixarene is slightly larger (13.4 Å) than the distance to the adjacent methyl groups in the other calixarene (12.3 Å). In spite of the similar distance (13.4 Å vs. 12.3 Å), the different orientation of the urea residues of the two calixarenes could allow a selective connection of functional groups attached to the same calixarene. In the following we will name such a reaction a-connection to distiguish it from a b-connection taking place between functional groups belonging to different calixarenes. 5.3.2 First Attempts

The oleﬁn metathesis between alkenyl groups has been frequently used in recent years to stabilize molecules (e.g., dendrimers [46]) or molecular assemblies [47] by additional covalent connectivities [48] and to synthesize macrocycles or more demanding structures like multiple catenanes [24]. Therefore, this reaction was chosen for the connection of alkenyl residues attached to a calix[4]arene via the urea residues. To avoid complications by cis/trans isomerism around the newly formed double bond, the crude reaction mixture was always hydrogenated before working up (Scheme 5.10). In a ﬁrst attempt, we tried to connect the alkenyl residues attached to the urea groups within a homodimer of 5. A mixture of three products was obtained and separated chromatographically. Mass spectrometry revealed the expected molecular mass, while their symmetry, indicated for instance by the number of low-ﬁeld NH signals

5.4 Multimacrocycles

2

O-(CH2)m-CH=CH2

O-(CH2)m-CH=CH-(CH2)m- O

+

CH2=CH2

H2 O-(CH2)n- O

n = 2•m + 2

Scheme 5.10 Connection of pendant alkenyl residues by oleﬁn metathesis, followed by exhaustive hydrogenation of the newly formed double bonds.

Scheme 5.11 Reaction products obtained by intramolecular oleﬁn metathesis of the hydrogen bonded dimer 55, identiﬁed by 1 H NMR and ESI-MS spectra.

in 1H NMR spectra, led to their identiﬁcation. While the starting homodimer shows only one NH signal (at 10 ppm), four signals with equal intensity correspond to the envisaged bis[2]catenane formed by four a-connections. A doubly bridged [2]catenane (eight NH signals) is formed by two a- and two b-connections and a tetra-bridged capsule (two NH signals) by four b-connections [49] (Scheme 5.11). Considering the distances discussed above (Figure 5.6), this outcome is not entirely surprising. To obtain selectively a single, well-deﬁned product, it is obviously not sufﬁcient to arrange the reacting groups in an appropriate mutual position. A perfect preorganization also demands the separation of those functional groups which should not be involved in the reaction. For an intramolecular connection between reactive functions attached to the urea residues, this preorganization is possible in heterodimers with a non-reactive tetra-urea calix[4]arene.

5.4 Multimacrocycles 5.4.1 Template Synthesis of Multimacrocyclic Calix[4]arenes

The intramolecular connection of two functional groups attached to the wide rim of calix[4]arene can be achieved under high-dilution conditions [50]. If the number of functional groups increases, a selective connection becomes more and more difﬁcult, as illustrated in Figure 5.7. For methyl ethers, the situation is even more complicated by the presence of several conformations.

j155

156

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes R YO YO OY

YO YO OY OY

YO YO OY OY OY

R

a

R b R a Y C3H7

R

R

R

R

R

Y = CH3

R R

RR

R R

R

Y C3H7

Figure 5.7 Possible side reactions and complications for the controlled intramolecular connection of functional groups R attached to the wide rim of a calix[4]arene.

Usual high-dilution conditions will not prevent a connection of opposite groups in 5, and the controlled connection of eight alkenyl groups in 6 seems impossible (see Figure 5.8). In heterodimers with tetra-tosylurea 4 (Figure 5.4), however, the alkenyl residues are arranged in a way, which allows their interaction with the immediate neighbors only.3) Furthermore, the template 4 induces the cone-conformation for tetramethylethers of type 5 or 6, while tetramethoxy calix[4]arenes usually prefer the partial cone conformation. In practice it is reasonable to use the non-reactive compound in the heterodimer, the tetra-tosylurea 4, in a small excess, to exclude the undesired formation of homodimers 55 or 66. After metathesis, the dimer is decomposed by a hydrogen-bond breaking solvent (usually THF), and the crude product is usually hydrogenated directly, to avoid complications by cis-trans isomerism. The whole reaction sequence is schematically summarized in Scheme 5.12. Bis- and tetraloop calix[4]arenes 8 and 9 were easily prepared in high yields (70–95%) [51,52] for residues Y ¼methyl, pentyl, and dodecyl, and loop sizes of n ¼ 2m þ 2 ¼ 8, 10, 14, and 20.4) Several aspects show the generality of this approach: a. The structure of the chains may be varied, using various different tetra-urea derivatives. Up to now, we have realized combinations of oligomethylene with oligoethyleneoxy chains [53]. b. As shown in Scheme 5.12, there is also a rational possibility of converting an AABB-type tetra-urea 7 into a trisloop compound 10, possibly with different length of the connecting chain(s) [54]. Even in this case, the yield of pure 10 is surprisingly high (70–75%), although one undesired connection is possible between the two monoalkenyl urea residues. c. Tetraloop compounds with different size and structure of the loops (characterized by X, Y, Z) are accessible; e.g. AAAB-type 3) A transcavity contact should be possible only for

very long chains.

4) Up to m = 11, leading to n = 20 there was no

indication of a reaction between two alkenyl groups attached to the same phenyl ring.

5.4 Multimacrocycles Y

Y

Y

NH

NH

4

O C

X

X

5

2

NH

X O

O

NH O C

NH

NH

NH

O

NH

4

O C

O C

Y

O

O

O

O

O

O

O

X1

X2

X2

7

6

a

X = (CH2)mCH=CH2

a

X = (CH2)mCH=CH2

b

X = (CH2CH2O)pCH2CH=CH2

b

X = (CH2CH2O)pCH2CH=CH2

X1, X2 = (CH2)mCH=CH2

Figure 5.8 Formula survey of tetra-urea calix[4]arenes used as precursors for bis-, tris-, and tetraloop tetra-ureas (Y ¼ alkyl).

Tos

Tos

Tos

Tos

Tos

Tos

a

Tos

Tos

8·4

b

Tos

7·4

a

a

Tos

Tos

Tos

Tos

Tos

9·4

b

8

Tos

6·4

5·4

Tos

Tos

Tos

Tos

Tos

Tos

Tos

Tos

Tos

10·4

b

9

Scheme 5.12 Selective synthesis of bis-, tris-, and tetraloop tetra-urea calix[4]arenes: (a) metathesis and subsequent hydrogenation, (b) decomposition of the heterodimer (a pseudorotaxane) and isolation of the multiloop calixarene.

10

j157

158

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

O N C

a

H

X

a

a

O

H

O

O

O N

H

N C

N H

O

O

O

N C

a

H

H

O H

N

O

O N

N C

H

O

X O

a

Y

H C N N

O

N H O

H O

O H

C N

b O

AAAB-type

O

b

C N N

O

a

H N H O

H

H

C N O

O

O

Y X a

a

O O

a

N C H

O

H

O

N C

N H

O

O

O N

H

O

N C

a

H

H

O H

N

O N

N C

H

O

O

Y O

b

C N N

O

H O

Y

H

b

N H O

O H

C N

b

O

C N N

O

b

H

O

AABB-type

N H O

H

H

C N

O

O O

Z Figure 5.9 Approaches to tetraloop compounds with different size and structure of the loops via heterodimerization with tetratosylurea 4 followed by oleﬁn metathesis and hydrogenation. X ¼ a-(CH2)2-a; Y ¼ a-(CH2)2-b; Z ¼ b-(CH2)2-b.

ureas (compare Figure 5.5) lead to two different loops with the sequence XXYY, while the AABB-type gives three different loops with the sequence XYZY (see examples in Figure 5.9). d. Finally, the reaction forming the macrocycle is not restricted to oleﬁn metathesis. In principle, all reactions which proceed under conditions where the hydrogen bonded dimers exist should be suitable. (For an example, see Section 5.5) The structure of bis- to tetraloop compounds is clearly proved by their ESI mass spectra (which usually show the molecular ion with high abundance) and by their 1 H NMR spectra in hydrogen bond breaking solvents, where they exist as monomeric species. Examples of symmetrical bis- and tetraloop compounds are shown in Figure 5.10b,c. They demonstrate simultaneously that the molecules are kept in the

Figure 5.10 Molecular conformation of a tetraloop calix[4]arene 9 (Y¼Me,n¼10) in the crystalline state (a). 1H NMR spectra (400 MHz) of a bisloop compound 8 (Y¼Me,n¼10) in DMSO-d6 (H2O peak is marked by asterisk) at 25 C (b) and of a tetraloop compound

9 (Y¼Me,n¼14) in pyridine-d6 (marked by asterisks) at 100 C (c). A hydrogen bond-breaking solvent and elevated temperature are necessary to obtain a well-resolved spectrum of the monomeric molecule.

5.4 Multimacrocycles

j159

160

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes cone conformation, although tetramethylethers usually prefer the partial cone conformation. It may be assumed that a conformational interconversion becomes possible again for longer connecting chains. Indications are found already for bisloop compounds with n ¼ 20 [53]. 1 H NMR spectra in apolar solvents are less informative. Broad unresolved signals indicate the presence of undeﬁned species. Obviously, the multicyclic structure does not allow the formation of regular dimers, which would lead to an unfavorable overlap of the loops, but the interaction via the urea functions creates irregular aggregates. This inability to form hydrogen-bonded homodimers can be exploited in further reactions, e.g., for the construction of multiple catenanes (see Section 5.5). A single-crystal X-ray structure for a tetraloop compound shows the molecules (not unexpectedly) in a pinched cone conformation (see Figure 5.10a), although the time-averaged conformation deduced from the 1 H NMR-spectra is C4v-symmetrical (Figure 5.10b). 5.4.2 Double Template Synthesis of Giant Macrocycles

Tetraloop compounds 9 contain a huge macrocycle (formed by fourfold metathesis), which is covalently connected to the calix[4]arene skeleton via four urea linkages. Cleavage of these connections is easily possible in boiling acetic acid (Scheme 5.13). Thus, ﬂexible, macrocyclic molecules 11 are available in 50–75% yield [53], which would be difﬁcult to obtain by non-templated macrocyclization reactions under high-dilution conditions. The four N-(3,5-dioxyphenyl)acetamide fragments of the macrocyle 11 may be connected not only by oligomethylene chains of various lengths (up to now n 20, corresponding to a total ring size of 100 atoms). Various other structural elements can be incorporated, and connecting chains may also be different in size and

O H

O

N C H

H

N H

(CH2)n

C N N

O

H

O

N C

Y O OY YO O Y

O

(CH2)n O O

O

H N

N H

N H

O O

H

O

O (CH2)n

(CH2)n

O

reflux, 12h

H

C N

9

H N

N H

O

O O

(CH2)n

N

O

CH3CO2H

H

H

O

O

O

(CH2)n

N

H

O

N

N Y O OY YO O Y

O

O

O

H N

(CH2)n

O

O

O

O (CH2)n

Scheme 5.13 Release of a macrocyclic molecule 11 by cleavage of the urea groups.

11

5.4 Multimacrocycles

Tos

1)

I

2)

I

Tos

II

Tos

Tos Tos

I

Tos

II

3) Tos

Tos

Tos

5)

11

I

Tos

4)

9

II

Tos

Tos

Scheme 5.14 Reaction sequence (schematic) for the synthesis of huge macrocycles using a calix[4]arene twice as template. The single steps are discussed in the text.

structure. From these possibilities, macrocyles containing ethyleneoxide segments have been realized so far [53]. The total reaction sequence, in which a macrocycle is formed in a well-deﬁned way by the connection of four reactive, bifunctional fragments, involves two calix[4] arenes as template, as schematically shown in Scheme 5.14 and described below: 1. First a tetraamino calix[4]arene (template I) is used to attach four a,o-bisethylene fragments via urea groups. Thus, the number (four) and, as indicated above, eventually the sequence of the fragments to be connected, is determined. 2. A second calix[4]arene, the tetra-tosylurea (template II) is used to bring the reactive groups into the correct position for their connection. 3. Metathesis, followed by hydrogenation, converts the heterodimer into a pseudorotaxane. 4. After the reaction, a hydrogen bond breaking solvent liberates template II, which can be used again. 5. Finally, the urea links are cleaved to isolate the target product (the macrocycle) and template I (as tetraacetamide), which can in principle be used again. (In practice it is easier to prepare the tetraamino calix[4]arene according to Scheme 5.8.) It should be possible to synthesize still larger and more sophisticated macrocyclic molecules in a similar manner.

j161

162

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes 5.5 Multiple Catenanes Based on Calix[4]arenes 5.5.1 General Considerations

The inability of bis-, tris-, or tetraloop compounds to form homodimers and the general tendency of tetra-urea calix[4]arenes to dimerize can be further exploited for the synthesis of mechanically interlocked molecules. A 1:1 mixture of tetra-ureas 5 or 6 with bis- or tetraloop compounds 8 or 9 (in practice the non-reactive loop component is added in a small excess) contains exclusively heterodimers (e.g., 58, 59, or 69), since this is the only possibility, to have all the urea functions involved in the favorable hydrogen-bonded belt.5) Again this is easily evidenced by the complete absence of peaks for the homodimer 55, or 66 in the 1 H NMR spectra (for an example see Figure 5.11). Like in heterodimers with tetra-tosylurea (4), the alkenyl groups of 5 and 6 are prearranged in these heterodimers with 8 or 9 in a way which allows only one welldeﬁned connection by metathesis reaction. The only difference is that this now leads

* OM e

C6H3 C6H3 C6H2

NH

C6H2

NH

ArCH2Ar

6·6

ArCH2Ar

C6H3

*

C6H3 C6H2 NH

C6H3

9.0

C6H2

C6H3

6·9

OM e ArCH2Ar ArCH2Ar

NH

8.0

7.0

6.0

5.0

4.0

3.0 δ/ppm

Figure 5.11 1H NMR (400 MHz, 25 C) spectroscopic evidence for the exclusive formation of heterodimers in a stoichiometric mixture of 6 (Y¼Me,n¼6) and 9 (Y¼Me,n¼14) in CD2Cl2 (marked by asterisk). Mainly the region between 6.3 and 7.0 ppm shows that no signals for the homodimer 66 are present.

5) A similar principle was used in the Cu(I)

templated synthesis of various catenanes

(see Sauvage and coworkers in Refs. [9–11]).

5.5 Multiple Catenanes Based on Calix[4]arenes

5·8

12

5·9

6·9

13

14

Scheme 5.15 Controlled syntheses of bis[2]catenanes 12, bis[3]catenanes 13 and cyclo [8]catenanes 14.

to catenated ring structures. Thus, it is not entirely surprising that bis[2]catenanes 12 (from 58), bis[3]catenanes 13 (from 59 or 68) and cyclic [8]catenanes 14 (from 69) are formed in excellent yields of 50–97% (isolated and puriﬁed product), as schematically represented in Scheme 5.15 [55]. The formation of eight interlocked ring-systems in one step and in a yield of up to 90% is a particularly striking example of the effect of preorganization. Based on all experience, it is justiﬁed to predict, that the combination of trisloop compounds with bis- or tetraloop compounds to give the respective catenanes will also proceed selectively with high yield (Scheme 5.16), since only one heterodimer is possible in all these cases. Only the combination of a trisloop compound 10 with its alkenyl precursor 7 may lead to two regioisomeric dimers. 5.5.2 Bis[2]catenanes

The synthesis of bis[2]catenanes outlined in Scheme 5.15 allows not only the preparation of compounds with identical loops (12a). The rings attached to the second bisloop calix[4]arene may be also different from those of the ﬁrst calix[4]arene (12b) [55]. Two regioisomeric bis[2]catenanes are possible, where the two rings attached to a given calix[4]arene are different (Figure 5.12) The isomer 12c can be selectively synthesized from a bisalkenyl urea, where the other two urea functions are connected to a loop, while the third regioisomer 12d would be not so easily available. The rational synthesis of 12c is shown in Scheme 5.17. The starting monoloop di-urea 15 is prepared in the conventional way analogous to Scheme 5.9 by macrocyclization of the di-Boc derivative with an activated bisurethane under dilution conditions, followed by deprotection and acylation [50]. Although 15 belongs to the

j163

164

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

5·10

7·8

6·10

7·9

Scheme 5.16 Possible synthetic pathways to calix[4]arene-based multiple catenanes involving trisloop derivatives.

12b

12c

12d

Figure 5.12 Regioisomeric bis[2]catenanes with pairwise different rings. (CH2)8 O

O

H

α

O

N C H

H H

N

N

N C

O

2 C NH N H

O

15

(

H

β

C NH

β′

15·15

O (

(

O m

N

m = 6, 9

(

O

m

Scheme 5.17 Synthesis of bis[2]catenanes of type 12c. Instead of the desired a-connections between alkenyl residues belonging to the same calixarene molecule, a b-connection between the two molecules of the dimer 1515 is also possible. This would leave two remote alkenyl groups for a potential b0 -connection.

12c

5.5 Multiple Catenanes Based on Calix[4]arenes

AABB-type it forms only one regioisomeric dimer 1515, since the two macro-rings do not overlap for steric reasons. In dimer 1515, there is still one possibility for the undesired connection between (adjacent) alkenyl residues belonging to different calixarenes (b-connection), which may be the reason for the lower yield of 50%. For longer chains also a connection of remote alkenyl groups (b0 -connection) cannot be excluded (see Section 5.5.3). All bis[2]catenanes are chiral. Figure 5.13 illustrates their symmetry properties. 1 H NMR spectra are in agreement with the predicted symmetry, although for a C1symmetric species, such as 12c, or a mixture of two diastereomeric C2-symmetrical compounds of type 12b a complete resolution of all peaks is usually not possible. The optical resolution was achieved in most cases on an analytical scale by chromatography [52] (see the example in Figure 5.14) using chiral stationary phases of the Chiralcel OD or Chiralpak AD-H type. The separation factors vary between close to 1 and 2.36. In two cases, it was possible to conﬁrm the structure by single-crystal X-ray analysis, although the atoms of the interlocking chains are usually strongly disordered. Figure 5.15 shows one example. The crystal was grown from chloroform/ methanol, a hydrogen bond breaking mixture, and the dimeric capsule is completely opened. The whole assembly is connected only by the interlocked rings. This disrupted capsule suggests that a guest exchange should be easily possible in bis[2]catenanes already for rings with n ¼ 10.

12b

12a

D2

C2

12c

C2

C2

C2

C2

C2

C2

C2

Figure 5.13 Symmetry properties of bis[2]catenanes with identical (12a) and different rings (12b and 12c).

C1

C1

j165

166

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

Figure 5.14 Example of the chromatographic resolution of 12a. Column: Chiralcel OD, eluent: hexane/ethanol (90/10, v/v), CDspectra of both fractions: hexane/ethanol (90/10, v/v), cell length: 25 mm.

5.5.3 Towards Novel Topologies

When the synthesis outlined in Scheme 5.17 was applied to larger loops (e.g., n ¼ 20), the puriﬁcation of the product became more difﬁcult or even impossible. While the samples showed the expected ESI mass spectrum, a satisfactory puriﬁcation from a chromatographically similar compound was not possible, and the 1 H NMR spectrum was not entirely in agreement with the expected structure. This led

Figure 5.15 Molecular structure of 12a as determined by X-ray analysis (Y ¼ pentyl, n ¼ 10). Ether residues, solvent, and H-atoms are omitted for clarity.

5.5 Multiple Catenanes Based on Calix[4]arenes O N H C N H

O

O

H N H N C

() p

O

(CH2)n O

N

16

O

H

C NH

`

O

C NH N H

β

16

16·16

17

0

Scheme 5.18 Probing the possibility of a b -connection: Dimerization and metathesis reaction of tetra-urea 16 (n ¼ 8, 10; p ¼ 9).

to the idea that an isomeric compound was formed by a combination of a b- and a b0 -connection. Molecular models suggest that such a b0 -connection between remote alkenyl residues in a dimer is possible. To check whether it really occurs under the conditions of the metathesis reaction, we synthesized the monoloop tetra-urea compounds 16, in which one of the non-cyclic urea residues is substituted by a bulky group which cannot penetrate the loop (Scheme 5.18). (The synthesis is analogous to that of 15, introducing in the ﬁnal two steps the bulky residue ﬁrst.). Formally compound 16 belongs to the AABC-type of tetra-ureas (compare Section 5.3.2) for which numerous regioisomeric dimers are possible. However, these substituents at the urea residues ensure that selectively only one dimer is formed, in which no overlap of the loops occurs and no penetration of the loop by the bulky group takes place. 1 H NMR spectra are in agreement with the formation of a single C2-symmetrical dimer, which is necessarily composed of the same enantiomer of 16. In fact, after metathesis and hydrogenation a single reaction product is isolated in 40–55% yield, which cannot be split in its calix[4]arenes by hydrogen bond breaking solvents [56]. Although the structure for several single crystals could not be solved, the 1 H NMR and ESI mass spectra serve as an unambiguous structural proof. The topology of these molecules becomes more evident by the transformations shown in Figure 5.16. The representation reached after some simpliﬁcations, which

j167

168

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes (b)

(a)

(c)

Figure 5.16 Illustration of the topology of 17 by stepwise simpliﬁcation: (a) omission of non-cyclic urea arms; (b) graphical reduction of each calixarene to a branching point; (c) optical rearrangement of the graph.

do not change the topology, shows, that there are structural elements of catenanes and rotaxanes (and even some similarities to knots).6) 5.5.4 Bis[3]catenanes and Cyclic [8]Catenanes

In contrast to bis[2]catenanes 12, which are topologically chiral, bis[3]catenanes 13 and cyclic [8]catenanes 14 prepared via metathesis and subsequent hydrogenation are usually not chiral. For identical loops, their symmetry (without and with the directionality of the hydrogen-bonded belt) is C2v/C2 and D4d/S8, respectively. This is reﬂected by their NMR-spectra (for an example, see Figure 5.17). The structure of 14 (n ¼ 14) was unambiguously proved by single-crystal X-ray analysis (Figure 5.18). The space-ﬁlling representation suggests a more or less

C6H2 NH

C6H3 C6H3 C6H3

14

ArCH2Ar

C6H2

ArCH2Ar NH

* 9.0

7.0

5.0

Figure 5.17 H NMR spectrum (400 MHz, 25 C) of a cyclic [8]catenane of type 14 (Y ¼ methyl, n ¼ 14) in CD2Cl2 (marked by asterisk) reﬂecting S8-symmetry of the molecule (compare Fig. 5.1b). 1

6) For a similar rotaxane, see Ref. [3].

3.0

δ/ppm

5.5 Multiple Catenanes Based on Calix[4]arenes

Figure 5.18 Molecular shape of the cyclic [8]catenane 14 (Y ¼ methyl, n ¼ 14) in the crystalline state, showing the interlocking of eight rings. Space-ﬁlling representation seen along the S8-axis (a) and perpendicular to it (b). Alkyl residues and hydrogen atoms are omitted for clarity.

complete shielding of the included guest. However, it was shown by ESI mass spectra that, for the examples studied so far, an exchange of the guest is still possible. A solution of 13 or 14 in a 1:1:0.4 mixture of 1,4-diﬂuorobenzene, dichloromethane, and methanol showed a strong peak corresponding to the doubly charged species of the respective catenane 13 or 14 þ 2Naþ þ C6H4F2. Since 1,4-diﬂuorobenzene was neither present during the synthesis nor during work-up, its inclusion during the MS-experiments with the catenanes remains the only reasonable explanation. As discussed before, the strategy to synthesize catenanes is not necessarily restricted to oleﬁn metathesis as the ring closing reaction. Any reaction which can be carried out under conditions where the hydrogen-bonded dimers exist should be appropriate. A ﬁrst example was realized by the formation of anhydride linkages with dicyclohexylcarbodiimide as the reagent in benzene as the solvent [57], as schematically illustrated in Scheme 5.19. Y

H O O

O

NH

H O O

O

O

O

DCC

4

O C NH

18

Be nze ne O O( ) OH n

Y = C10H21; n = 4

18·9 O O H

O O H

O

O

19 O

Scheme 5.19 Synthesis of a bis[3]catenane 19 with two cyclic anhydride structures.

j169

170

FG

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes FG

a FG

b

FG

Scheme 5.20 Two possibilities to synthesize tetra[2]rotaxanes (schematic): (a) by stoppering (FG represents a suitable functional group) or (b) by clipping (ring closure by metathesis).

5.6 Multiple Rotaxanes

The preorganization of functional groups attached to the urea residues could also be used for the synthesis of multiple rotaxanes. Schematically, this is shown for tetra[2] rotaxanes in Scheme 5.20. Bis[2]rotaxanes will be available analogously, using a bisloop compound 8 instead of the tetraloop compound 9. In principle there are two possible pathways: . Bulky residues can be attached to the urea groups of a tetra-urea under conditions where it forms a heterodimer (a pseudorotaxane) with a bis-, tris-, or tetraloop tetra-urea. Often this strategy is called stoppering (see Scheme 5.7), because stoppers are attached to the axles. . Multiple ring-closure reactions between adjacent urea functions of a suitably functionalized tetra-urea in a heterodimer with a second tetra-urea bearing bulky groups can also create the structural elements of a rotaxane. This strategy is usually called clipping (see Scheme 5.7). To realize the ﬁrst possibility, the tetra-ureas 20 bearing maleic imide functions at all four urea arms were synthesized. These dienophiles react with anthracene derivatives, here 1,4,5,8-tetrapentoxyanthracene 21, in a Diels-Alder cycloaddition (Scheme 5.21). Tetra-ureas 20 form homodimers in solvents such as CDCl3 or C6D6, but they are completely converted to heterodimers upon addition of an equimolar amount of a tetra-loop compound 9. When a 5–10% excess of the anthracene derivative 21 was applied, a quantitative conversion was obtained in reﬂuxing toluene after 72 h (as judged by NMR), and the tetra[2]rotaxane 22 was isolated in 40–50% yield. The rotaxane structure of 22 can be demonstrated by ESI-MS or MALDI-TOF-MS. It remains intact in THF-d8, a solvent in which the tetra-urea dimers usually dissociate. However, an upﬁeld shift of the NH-signals in 1 H NMR indicates that the

5.6 Multiple Rotaxanes

j171

OR

OR

OR Y

21

O

O

O

N

N

RO

4

O

OR

RO

O

O NH

RO

RO

RO O

RO

O

OR

N

N

N

N

O

OR

O C NH O

O O

20

() n

O N

N

Diels-Alder cycloaddition

O

N O

O

20·9

OR

OR

OR

OR

O

O OR

OR

O

O

OR

22

Scheme 5.21 Formation of tetra[2]rotaxanes 22 (n ¼ 8, 10; Y ¼ R ¼ C5H11) by stoppering. Introduction of bulky stopper residues by Diels-Alder cycloaddition.

hydrogen-bonded belt is weakened and that the urea functions are partly solvated by THF [58]. The synthesis of fourfold [2]rotaxanes by clipping requires the efﬁcient formation of heterodimers between a tetra-urea substituted by bulky stopper groups and an octaalkenyl urea 6. We initially hoped that the steric crowding in homodimers of a tetratritylphenylurea calix[4]arene would be sufﬁcient to shift the equilibrium toward the heterodimers in a mixture with 6. However, the distribution of the dimers was close to the statistical ratio and the desired rotaxane could be obtained in only 5% yield [59]. On the other hand, when smaller rings were synthesized according to the template reaction discussed in Section 4 (Scheme 5.12), it became difﬁcult to split the tetraloop calix[4]arene 9 from the tetra-tosylurea 4 used as template. For n ¼ 6, a splitting was eventually completely impossible. This follows for instance from the fact that the guest (here CH2Cl2), necessarily included when the heterodimeric capsule is formed, is not released under hydrogen bond breaking conditions. Thus, the molecules 23 represent the ﬁrst examples of fourfold [2]rotaxanes which were obtained by a clipping reaction [60]. In spite of having grown several single crystals, it was not possible to solve the structure of 23. To illustrate the shape of the rotaxane, Figure 5.19 shows an energyminimized structure obtained by MD simulation. It is evident that the tetrahedral arrangement of the substituents around the sulfur atoms (N-S-C angle ca. 109 ) prevents a slipping of the loops. It is not surprising that the stability of 23 depends on the size of the loops. For n ¼ 8, slow dissociation occurs in THF (estimated half-life 10–20 days), while immediate decomposition occurs in DMSO or pyridine. For n ¼ 6, however, no dissociation was observed after several weeks. An exchange of the included guest(s) is possible also for the intact rotaxane. For n ¼ 8, ethyl acetate (included during the chromatographic isolation) is exchanged against the solvent THF-d8 in a ﬁrst-order reaction (t1/2 ¼ 32 min). For n ¼ 6, only one of two included dichloromethane molecules is lost (t1/2 12 days). This inclusion of two guest molecules in a dimeric capsule of the tetra-urea type was for the ﬁrst time unambiguously proved by

RO

172

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes

Figure 5.19 Molecular shape of rotaxane 23 (n ¼ 6), a heterodimeric assembly of a tetraloop tetra-urea 9 (stick representation) and tetra-tosylurea 4 (space ﬁlling), based on MD simulations. Ether groups (OY ¼ OC5H11,) are omitted in the formula.

ESI-MS. Obviously heterodimers of tetra-arylureas, e.g., tolylurea (1), with 4 have a larger internal volume than homodimers 11 (260–270 Å3 vs. 190–200 Å3, based on MD simulations) [61] because of differences in the hydrogen-bonding pattern [62].

5.7 Self-sorting and Formation of Larger Assemblies

The covalent connection of two tetra-urea calix[4]arenes via their narrow rims leads to molecules which are bifunctional with respect to their ability to form dimeric capsules. This was used by Rebek and Castellano to form self-assembled polycaps, including also alternating structures based on the heterodimerization of aryl- and tosylureas [42,63]. The idea was later extended to switchable polymers [64]. We were recently able to show that the covalent connection of two tetra-urea calix[4]arenes via a rigid spacer between urea functions also leads to molecules 24, which can form polymeric assemblies (Scheme 5.22) [57,65]. As shown in Section 5.5, bisloop or tetraloop tetra-ureas (8, 9) do not form homodimers, while monoloop compounds (e.g., 15) form selectively only one homodimer. These selectivities may be summarized as follows: 1. Tetra-urea molecules in which adjacent urea functions are covalently connected form only those dimers in which no overlapping of the loops occurs. 2. If urea functions with bulky substituents are present, only those dimers can be formed in which the bulky residues need not thread through a loop. With the kind and size of loops studied so far, no exception to the ﬁrst rule was found, while the situation is less clear for the second rule. The threading of a larger

5.7 Self-sorting and Formation of Larger Assemblies

Y

Y Y Y O O O O

n

NH NH HN HN O OC C CO NH HN NH HN R R R X

O

j173

C

R R R NH HN HN NH CO C OC C O HN NH HN O NH

O O O O Y

Y Y

Y

24

X

N N

Scheme 5.22 Tetra-urea calix[4]arenes 24 connected via a rigid spacer X between their urea functions, and schematic representation of their self-organization.

urea residue may be sometimes a kinetic problem, depending also on the reaction conditions (solvent, temperature, etc.). In combination with the tendency to form dimeric capsules in apolar solvents, these rules describe a self-sorting process, illustrated in Scheme 5.23 by one example. The principle has already been used for the synthesis of multicatenanes (Section 5.5) and the synthesis of rotaxanes by stoppering (Section 6). Such a self-sorting process can also lead to the formation of structurally welldeﬁned assemblies of the dendrimer type [66]. We have demonstrated this in a ﬁrst example, using the dimerization of tri-ureas 26 derived from triphenylmethanes [67]. It has been shown that this dimerization is independent of the dimerization of tetra-urea calix[4]arenes. It is neither disturbed by tetra-ureas nor does it disturb their dimerization. In addition, Figure 5.20 even shows that the dimerization of rigidiﬁed tetra-ureas 25 also occurs independently [68]. Therefore, we connected the tri-urea 27 with three tetra-ureas 28, via covalent linkers, to a building block A. In apolar solvents molecules A alone would form crosslinked polymers via dimeric substructures formed by the tri- and tetra-ureas. Together with the building block B, a tetra-tosylurea substituted by four phthalimido alkylether residues, this polyassociation is prevented by the formation of heterodimers between tetra-aryl- and tetra-tosylureas. A solution of A and B in the ratio 2: 6 therefore contains exclusively the dendritic assembly shown in Figure 5.21, as proved by light scattering and DOSY NMR spectroscopy [69]. The self-assembled dendrimer carries 24 phthalimide groups (= protected amino groups) on its surface, which can be replaced also by other groups (e.g., dyes). Even larger structures should be available, using the whole spectrum of selectivities indicated above.

174

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes Tol

Tol Tol Tol

Tol

Tol

+

Tol

Tol

Tol

15

15·15 Tol

Tol

Tol

+

15·15 +

Tol Tol

Tol

Tol

Tol

+ Tol

Tol

Tol

Tol

Tol

Tol

1·15

Tol Tol

Tol

8

Tol

1

1·8 Tol

Tol Tol

Tol Tol

1·1

Scheme 5.23 Self-sorting of open (1), monoloop (15 with tolyl urea residues), and bisloop (8) tetra-ureas. The addition of bisloop compound 8 converts the mixture of three dimers 11, 115, and 1515 into a mixture of only two dimers, 18 and 1515, the only possibility to have all tetra-ureas assembled in a dimer.

C5H11

O

O

1

NH O C

O

*

**

NH O C NH NH

NH O C

NH

*

O

CH

O

25 4

EtO O

26 2

NH O C

3

NH

*

* ** *

9.0

8.5 1

6.5

6.0

Figure 5.20 Partial H NMR spectrum (400 MHz, CDCl3, 25 C) of a mixture of tetra-ureas 1 and 25 and tri-urea 26 showing only signals of the three homodimers. This proves the independent dimerization of the three different tri/tetra-urea molecules.

δ /ppm

B

Figure 5.21 Self-assembly of two building blocks A and six building blocks B to a structurally uniform dendrimer.

A

O

Y =

OH

NH

O

O OH

H O O

NH O C C O NH O C NH NH CH3

NH

N

O

O

HO

O

27

H2N

S

Y

O

NH O S O

O

Y

NH NH O C C O NH NH O O O S O S

Y

NH O C C O NH NH

O

NH

O O O

NH NH O C

Y

NH

NH NH O C

NH

C

NH

C

O

O

O

O O O

B

28

5.7 Self-sorting and Formation of Larger Assemblies

j175

176

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes 5.8 Conclusions and Outlook

The principal idea of this present essay was to show how the unique preorganization of functional groups in self-assembled dimers of tetra-urea calix[4]arenes can be used to prepare novel multi-rotaxanes and -catenanes or topologically even more complex molecules and supramolecular structures. We will conclude by summarizing some related studies in which calixarenes were used in a different way as building blocks for the construction of such structures or assemblies. As early as 1993, Gutsche and Kanamathareddy found that calix[6]arenes in which opposite oxygens are connected by a p-xylylene ether bridge have the tendency to assume a conformation which they dubbed self-anchored rotaxane [70,71]. However, the use of a calix[6]arene as a wheel for [2]rotaxanes did not occur before 2000 [72]. Their synthesis is based on the stability of pseudorotaxanes formed by 1,3,5triphenylureidocalix[6]arenes and dialkylviologen cations as an axle [73], which allows the introduction of stoppers by esteriﬁcation of the hydroxyl end groups of the axle (reﬂux with diphenyl acetyl chloride in toluene for 72 h). Since the threading of an asymmetrical dipyridinium axle occurs exclusively from the wide rim [74], it was even possible to prepare constitutionally isomeric rotaxanes, using two different stoppers [75]. In view of these examples, it is astonishing that the larger calix[8]arene was not yet used as a macrocyclic component in rotaxanes or catenanes. Calix[4] arenes have been used, however, as stoppers in rotaxanes [76], where they can contribute to the binding of guests [77]. Examples of the incorporation of calixarenes into [2]catenanes are rare. They are not based on the calixarene as such as a macrocycle, but as a fragment for the construction of a cyclobis(paraquat-p-phenylene) [78,79]. Thus, with bis-p-phenylene-[34]crown-10 as the second ring (Figure 5.22a) a [2]catenane is prepared, as described at the beginning of this chapter, using donor-acceptor interactions for the necessary preorganization. The construction principle has recently been inverted, incorporating one of the phenolic units of a calix[4]arene in the crown ether ring (Figure 5.22b) [80].

a

+ N

N+

O O O O O O +N O O O O H H O OO O

b

R

R

R

O

O

O

+ O N O O OO O Y Y Y O

N+

O

+N

N+ O

O

4 Cl 4 Cl -

R

R

R

R

Figure 5.22 Examples of [2]catenanes based on the principle outlined in Scheme 5.6.

N +

5.9 Experimental: Selected Procedures

While these examples again demonstrate the unlimited possibilities for the chemical modiﬁcation of calixarenes [81], the preorganization obtained by the dimerization of tetra-urea derivatives opens the door to the construction of molecules or molecular assemblies with entirely novel topologies.

5.9 Experimental: Selected Procedures 5.9.1 Synthesis of Tetra-urea 5 (Y ¼ C5H11; m ¼ 9) [52]

A solution of tetraamino calix[4]arene (0.363 g, 0.474 mmol), 4-nitrophenyl-N-(3-(10undecenyloxy)phenyl)carbamate (1.0 g, 2.37 mmol), and triethylamine (0.343 mL, 2.37 mmol) in DMF (20 mL) was stirred for two days at room temperature. The reaction mixture was diluted by dichloromethane (250 mL), and a solution of K2CO3 (1 M, 150 mL) was added. The mixture was stirred for 1 h at room temperature, and the organic layer was washed with a solution of K2CO3 (0.3 M, 3 times 150 mL). After drying over MgSO4, the solvent was evaporated, and the desired tetra-urea was puriﬁed by column chromatography (silica gel, THF/hexane, 1:6). Reprecipitation from methanol/dichloromethane gave the desired tetra-urea 5 (Y ¼ C5H11; m ¼ 9) (0.617 g, 68% yield) as a colorless powder. 5.9.2 Synthesis of Bisloop Tetra-urea 8 (Y ¼ C5H11; n ¼ 20) [52]

A mixture of tetra-tosylurea 4 (0.233 g, 0.150 mmol) and tetra-urea 5 (Y ¼ C5H11; m ¼ 9) (0.250 g, 0.131 mmol) in benzene (265 mL) was heated under reﬂux for 5–10 min. During this time a clear solution was usually formed. After cooling to room temperature the complete formation of the heterodimer was checked by 1 H NMR (benzene-d6). The reaction mixture was purged with nitrogen for 30 min, and a solution of the Grubbs catalyst (0.022 g, 0.0261 mmol) in benzene (3 mL) was added in one portion. After two days of stirring at 20 C, triethylamine (1.5 mL) was added to destroy the catalyst and the complex with tetra-tosylurea 4. After 1 h, the solvent was evaporated and the residue redissolved in THF (20 mL). Hydrogenation of the crude product obtained by evaporation was carried out in THF (20 mL) with hydrogen (normal pressure) in the presence of platinum dioxide (83% Pt, 0.0266 g, 0.117 mmol). The solvent was evaporated, and the residue was passed through a column (silica gel, THF/hexane, 1:2). Reprecipitation from THF/hexane gave the desired product 8 (Y ¼ C5H11; n ¼ 20) (0.203 g, 84% yield) as a colorless powder. 5.9.3 Synthesis of Bis[2]catenane 12 (Y ¼ C5H11; n ¼ 20) [52]

A mixture of tetra-urea 5 (Y ¼ C5H11; m ¼ 9) (0.050 g, 0.0261 mmol) and bisloop compound 8 (Y ¼ C5H11; n ¼ 20) (0.0510 g, 0.0274 mmol) in benzene (60 mL) was

j177

178

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes heated under reﬂux until a clear solution was formed, which usually required 10–60 min. After cooling to room temperature, the formation of the heterodimer was checked by 1 H NMR (benzene-d6). The reaction mixture was purged with nitrogen for 30 min, and a solution of Grubbs catalyst (0.0043 g, 0.005 mmol) in benzene (5 mL) was added in one portion. After stirring at 20 C for 48 h, triethylamine (1 mL) was added and the mixture was stirred for 1 h. The solvent was evaporated, and the residue was dissolved in THF (20 mL) and hydrogenated for 2 h using platinum dioxide (83% Pt, 0.0053 g, 0.0235 mmol) as catalyst. The crude product obtained by evaporation was passed through a column (silica gel, THF/ hexane, 1:2). Recrystallization from chloroform/methanol gave the bis[2]catenane 12 (0.0854 g, 88% yield) as a colorless powder. 5.9.4 Synthesis of Tetra-urea 6a (Y ¼ C5H11; m ¼ 6) [53,55]

A solution of 3,5-di-(7-octenyloxy)benzoic acid (1.47 g, 3.92 mmol), diphenylphosphorylazide (1.19 ml, 4.31 mmol), and triethylamine (0.436 ml, 4.31 mmol) in toluene (79 ml) was stirred for 2 h at 55 C. Tetraamino calix[4]arene (0.500 g, 0.654 mmol) was added, and the mixture was stirred for 2 h at 85 C. Then the solvent was evaporated, and the residue was triturated with methanol. The crude product thus obtained was puriﬁed by column chromatography (silica gel, ethyl acetate/hexane, 1:20) followed by reprecipitation from methanol/dichloromethane, which gave the desired tetra-urea 6 (Y ¼ C5H11; m ¼ 6) (0.714 g, 48% yield) as a colorless powder. 5.9.5 Synthesis of Tetraloop Tetra-urea 9 (Y ¼ C5H11; n ¼ 14) [53,55]

Tetra-tosylurea 4 (0.311 g, 0.200 mmol) and tetra-urea 6 (Y ¼ C5H11; m ¼ 6) (0.300 g, 0.133 mmol) were dissolved in dichloromethane (540 mL). The solution was stirred under nitrogen for 4 h at 55 C. The reaction mixture was purged with nitrogen for 30 min, a solution of Grubbs catalyst (0.0440 g, 0.0533 mmol) in dichloromethane (5 mL) was added, and the stirring was continued for 48 h at room temperature. Then triethylamine (2 mL) was added, and the solution was stirred for 1 h to destroy the catalyst and the complex with tetra-tosylurea 4. After evaporation at 60 C, the residue was dissolved in dichloromethane (5 mL), and passed through a column (silica gel, THF/hexane, 1:2) to remove the catalyst and tetra-tosylurea rests. The crude product was hydrogenated in THF (25 mL) for 24 h (room temperature, normal pressure) using platinum dioxide (83% Pt, 0.0545 g; 0.240 mmol) as catalyst. The residue obtained by evaporation was puriﬁed by column chromatography (silica gel, THF/hexane, 1:2). After reprecipitation from chloroform/methanol the desired tetraloop calix[4]arene 9 (Y ¼ C5H11; n ¼ 14) (0.267 g, 93% yield) was ﬁnally obtained as a colorless powder.

Abbreviations

5.9.6 Synthesis of [8]Catenane 14 (Y ¼ C5H11; n ¼ 14) [55]

A solution of tetraloop compound 9 (Y ¼ C5H11; n ¼ 14) (0.0500 g, 0.0233 mmol) and 6 (Y ¼ C5H11; m ¼ 6) (0.0499 g, 0.0222 mmol) in freshly distilled and degassed dichloromethane (90 mL) was stirred at room temperature for 48 h; the complete formation of the heterodimer was veriﬁed by 1 H NMR. When the signals of the homodimer 6 had disappeared, the reaction mixture was purged with nitrogen for 30 min, and a solution of Grubbs catalyst (0.007 g, 0.0088 mmol) in dichloromethane (5 mL) was added. After stirring for 48 h, triethylamine (0.5 mL) was added and the solution was stirred for 1 h and evaporated. The residue was dissolved in dichloromethane (5 mL) and passed through a column (silica gel, THF/hexane, 1:12). The solvent was evaporated, the crude product was dissolved in THF (20 mL), platinum dioxide (83% Pt, 0.009 g, 0.0399 mmol) was added to the solution, and the reaction mixture was stirred vigorously under a hydrogen atmosphere for 6 h at room temperature. The ﬁnal puriﬁcation by column chromatography (silica gel, THF/hexane, 1:12) followed by recrystallization from chloroform/hexane gave the octacatenane 14 (Y ¼ C5H11; n ¼ 14) (0.0394 g, 41% yield) as a colorless crystalline powder.

Abbreviations

Ar Boc CD CMPO DMF DMSO DOSY ESI-MS Et Grubbs catalyst MALDI-TOF-MS MD Me MS NMR THF Tol Tos UV

aryl tert-butyloxycarbonyl circular dichroism carbamoylmethylphosphine oxide N,N0 -dimethylformamide dimethyl sulfoxide diffusion-ordered spectrometry electrospray ionization mass spectrometry ethyl bis-(tricyclohexylphosphine)benzylidine ruthenium(IV) dichloride matrix assisted laser desorption ionization – time of ﬂight mass spectrometry molecular dynamics methyl mass spectrometry nuclear magnetic resonance tetrahydrofuran tolyl tosyl ultraviolet

j179

180

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes References 1 Sauvage, J.-P. and Dietrich-Buchecker, C. (1999) Molecular Catenanes, Rotaxanes and Knots, Wiley-VCH, Weinheim. 2 Adjami, D., Hess, K., Köhler, F., Näther, C., Oeckler, O., Simon, A., Yamamoto, C., Okamoto, Y. and Herges, R. (2006) Chem. Eur. J., 12, 5434–5445. 3 (a) Steuerman, D.W., Tseng, H.-R., Peters, A.J., Flood, A.H., Jeppesen, J.O., Nielsen, K.A., Stoddart, J.F. and Heath, J.R. (2004) Angew. Chem., 116, 6648– 6653; Angew. Chem. Int. Ed., 43, 6486–6491; (b) Sauvage, J.-P. (2005) Chem. Commun., 1507–1510; (c) Kay, E. R., Leigh, D.A. and Zerbetto, F. (2007) Angew. Chem., 119, 72–196; Angew. Chem. Int. Ed., 46, 72–191. 4 For the ﬁrst synthesis see: Chichak, K.S., Cantrill, S.J., Pease, A.R., Chiu, S.-H., Cave, G.W.V., Atwood, J.L. and Stoddart, J.F. (2004) Science, 304, 1308–1312. 5 Philp, D. and Stoddart, J.F. (1996) Angew. Chem., 108, 1242–1286; Angew. Chem. Int. Ed. Engl., 35, 1154–1196. 6 Schill, G. and Lüttringhaus, A. (1964) Angew. Chem., 76, 567–568; Angew. Chem. Int. Ed., 3, 546–547. 7 (a) Ünsal, Ö. and Godt, A. (1999) Chem. Eur. J., 5, 1728–1733; (b) Duda, S. and Godt, A. (2003) Eur. J. Org. Chem., 3412– 3420. 8 (a) Chambron, J.-C., Sauvage, J.-P., Mislow, K., De Cian, A. and Fischer, J. (2001) Chem. Eur. J., 7, 4085–4096; (b) Flamigni, L., Talarico, A.M., Chambron, J.-C., Heitz, V., Linke, M., Fujita, N. and Sauvage, J.-P. (2004) Chem. Eur. J., 10, 2689–2699. 9 (a) Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1989) Angew. Chem., 101, 192–194; (1989) Angew. Chem. Int. Ed. Engl., 28, 189–192; (b) DietrichBuchecker, C.O., Rapenne, G. and Sauvage, J.-P. (1997) Chem. Commun., 2053–2054.

10 (a) Dietrich-Buchecker, C.O., Sauvage, J.P. and Kintzinger, J.P. (1983) Tetrahedron Lett., 46, 5095–5098; (b) Dietrich-Buchecker, C.O., Sauvage, J.-P. and Kern, J.M. (1984) J. Am. Chem. Soc., 106, 3043–3045. 11 For a recent example see: Bäuerle, P., Ammann, M., Wilde, M., Götz G., Mena-Osteritz, E., Rang, A. and Schalley, C.A. (2007) Angew. Chem., 119, 367–372; Angew. Chem. Int. Ed., 46, 363–368. 12 Frey, J., Kraus, T., Heitz, V. and Sauvage, J.-P. (2005) Chem. Commun., 5310–5312. 13 (a) Sauvage, J.-P. and Ward, M. (1991) Inorg. Chem., 30, 3869–3874; (b) Leigh, D.A., Lusby, P.J., Teat, S.J., Wilson, A.J. and Wong, J.K.Y. (2001) Angew. Chem., 113, 1586–1591; Angew. Chem. Int. Ed., 40, 1538–1543; (c) Fuller, A.M., Leigh, D. A., Lusby, P.J., Oswald, I.D.H., Parsons, S. and Walker, D.B. (2004) Angew. Chem., 116, 4004–4008; Angew. Chem. Int. Ed., 43, 3914–3918. 14 (a) Cantrill, S.J., Chichak, K.S., Peters, A. J. and Stoddart, J.F. (2005) Acc. Chem. Res., 38, 1–9; (b) Pentecost, C.D., Peters, A.J., Chichak, K.S., Cave, G.W.V., Cantrill, S.J. and Stoddart, J.F. (2006) Angew. Chem., 118, 4205–4210; Angew. Chem. Int. Ed., 45, 4099–4104. 15 (a) Beer, P.D., Sambrook, M.R. and Curiel, D. (2006) Chem. Commun., 20, 2105–2117; (b) Sambrook, M.R., Beer, P. D., Lankshear, M.D., Ludlow, R.F. and Wisner, J.A. (2006) Org. Biomol. Chem., 4, 1529–1538. 16 Sambrook, M.R., Beer, P.D., Wisner, J. A., Paul, R.L. and Cowley, A.R. (2004) J. Am. Chem. Soc., 126, 15364–15365. 17 (a) Hunter, C.A. (1992) J. Am. Chem. Soc., 114, 5303–5311; (b) Vögtle, F., Meier, S. and Hoss, R. (1992) Angew. Chem., 104, 1628–1631; Angew. Chem., Int. Ed. Engl., 31, 1619–1622; (c) Brodesser, G., Güther, R., Hoss, R., Meier, S., Ottens-Hildebrandt, S.,

References

18

19

20

21

22

Schmitz, J. and Vögtle, F. (1993) Pure Appl. Chem., 65, 2325–2328; (d) Carver, F.J., Hunter, C.A. and Shannon, R.J. (1994) J. Chem. Soc., Chem. Commun., 1277–1280. (a) Leigh, D.A., Moody, K., Smart, J.P., Watson, K.J. and Slawin, A.M.Z. (1996) Angew. Chem., 108, 321–326; Angew. Chem. Int. Ed. Engl., 35, 306–310; (b) Kidd, T.J., Leigh, D.A. and Wilson, A.J. (1999) J. Am. Chem. Soc., 121, 1599– 1600; (c) Vögtle, F., Dunnwald, T. and Schmidt, T. (1996) Acc. Chem. Res., 29, 451–460. (a) Vögtle, F., Jäger, R., Händel, M., Ottens-Hildebrandt, S. and Schmidt, W. (1996) Synthesis, 353–356; (b) Asakawa, M., Brancato, G., Fanti, M., Leigh, D.A., Shimizu, T., Slawin, A.M.Z., Wong, J.K. Y., Zerbetto, F. and Zhang, S. (2002) J. Am. Chem. Soc., 124, 2939–2950. (a) Safarowsky, O., Nieger, M., Fröhlich, R. and Vögtle, F. (2000) Angew. Chem., 112, 1699–1701; Angew. Chem. Int. Ed., 39, 1616–1618; (b) Lukin, O., Müller, W. M., Müller, U., Kaufmann, A., Schmidt, C., Leszczynski, J. and Vögtle, F. (2003) Chem. Eur. J., 9, 3507–3517; (c) Passaniti, P., Ceroni, P., Balzani, V., Lukin, O., Yoneva, A. and Vögtle, F. (2006) Chem. Eur. J., 12, 5685–5690. (a) Ashton, P.R., Chrystal, E.J.T., Glink, P.T., Menzer, S., Schiavo, C., Stoddart, J. F., Tasker, P.A. and Williams, D.J. (1995) Angew. Chem., 107, 2001–2004; Angew. Chem. Int. Ed. Engl., 34, 1869– 1871; (b) Ashton, P.R., Campbell, P.J., Chrystal, E.J.T., Glink, P.T., Menzer, S., Philp, D., Spencer, N., Stoddart, J.F., Tasker, P.A. and Williams, D.J. (1995) Angew. Chem., 107, 1997–2001; Angew. Chem. Int. Ed. Engl., 34, 1865–1869. (a) Ashton, P.R., Glink, P.T., Stoddart, J. F., Tasker, P.A., White, A.J.P. and Williams, D.J. (1996) Chem. Eur. J., 2, 729–736; (b) Horn, M., Ihringer, J., Glink, P.T. and Stoddart, J.F. (2003) Chem. Eur. J., 9, 4046–4054; (c) Williams, A.R., Northrop, B.H., Chang, T.,

23

24

25

26

Stoddart, J.F., White, A.J.P. and Williams, D.J. (2006) Angew. Chem., 118, 6817–6821; Angew. Chem. Int. Ed., 45, 6665–6669. (a) Badjic´, J.D., Cantrill, S.J., Grubbs, R. H., Guidry, E.N., Orenes, R. and Stoddart, J.F. (2004) Angew. Chem., 116, 3335–3340; Angew. Chem. Int. Ed., 43, 3273–3278. (a) Zhu, X.-Z. and Chen, C.-F. (2006) Chem. Eur. J., 12, 5603–5609; (b) Zhu, X.Z. and Chen, C.-F. (2005) J. Am. Chem. Soc., 127, 13158–13159. (a) Ashton, P.R., Goodnow, T.T., Kaifer, A.E., Reddington, M.V., Slawin, A.M.Z., Spencer, N., Stoddart, J.F., Vicent, C. and Williams, D.J. (1989) Angew. Chem., 101, 1404–1408; Angew. Chem. Int. Ed. Engl., 28, 1396–1399; (b) Vögtle, F., Müller, W. M., Müller, U., Bauer, M. and Rissanen, K. (1993) Angew. Chem., 105, 1356–1358; Angew. Chem. Int. Ed. Engl., 32, 1295– 1297; (c) Gunter, M.J., Hockless, D.C.R., Johnston, M.R., Skelton, B.W. and White, A.H. (1994) J. Am. Chem. Soc., 116, 4810–4823; (d) Li, Z.-T. and Becher, J. (1996) Chem. Commun., 639–640; (e) Ashton, P.R., Ballardini, R., Balzani, V., Be4 lohradský, M., Gandolﬁ, M.T., Philp, D., Prodi, L., Raymo, F.M., Reddington, M.V., Spencer, N., Stoddart, J.F., Venturi, M. and Williams, D.J. (1996) J. Am. Chem. Soc., 118, 4931–4951; (f) Hamilton, D.G., Sanders, J.K.M., Davies, J.E., Clegg, W. and Teat, S.J. (1997) Chem. Commun., 897–898; (g) Try, A.C., Harding, M.M., Hamilton, D.G. and Sanders, J.K.M. (1998) Chem. Commun., 723–724. (a) Amabilino, D.B., Ashton, P.R., Reder, A.S., Spencer, N. and Stoddart, J.F. (1994) Angew. Chem., 106, 1316–1319; Angew. Chem. Int. Ed. Engl., 33, 1286– 1290; (b) Amabilino, D.B., Ashton, P.R., Boyd, S.E., Lee, J.Y., Menzer, S., Stoddart, J.F. and Williams, D.J. (1997) Angew. Chem., 109, 2160–2162; Angew. Chem. Int. Ed. Engl., 36, 2070–2072.

j181

182

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes 27 Ashton, P.R., Philp, D., Spencer, N. and Stoddart, J.F. (1991) J. Chem. Soc. Chem. Commun., 1677–1679. 28 Bogdan, A., Rudzevich, Y., Vysotsky, M. O. and Böhmer, V. (2006) Chem. Commun., 2941–2952. 29 Gutsche, C.D. and Iqbal, M. (1990) Org. Synth., 68, 234–237. 30 Saadioui, M., Shivanyuk, A., Böhmer, V. and Vogt, W. (1999) J. Org. Chem., 64, 3774–3777. 31 Rudzevich, Y., Rudzevich, V., Schollmeyer, D. and Böhmer, V. (2007) Org. Lett. 9, 957–960. 32 (a) Shimizu, K.D. and Rebek, J. Jr (1995) Proc. Natl. Acad. Sci. U.S.A., 92, 12403– 12407; (b) Hamann, B.C., Shimizu, K.D. and Rebek, J. Jr (1996) Angew. Chem., 108, 1425–1427; Angew. Chem. Int. Ed., 35, 1326–1329; (c) Castellano, R.K., Kim, B.H. and Rebek, J. Jr (1997) J. Am. Chem. Soc., 119, 12671–12672. 33 A similar dimerization was recently observed for thiocalix[4]arenes substituted by four CMPO-residues. (a) Kasyan, O., Kalchenko, V., Bolte, M. and Böhmer, V. (2006) Chem. Commun., 1932–1934; (b) Rudzevich, V., Kasyan, O., Kalchenko, V., Bolte, M., Schollmeyer, D. and Böhmer, V. unpublished. 34 Mogck, O., Paulus, E.F., Böhmer, V., Thondorf, I. and Vogt, W. (1996) Chem. Commun., 2533–2534. 35 Broda, F., Vysotsky, M.O., Böhmer, V. and Thondorf, I. (2006) Org. Biomol. Chem., 4, 2424–2432. 36 Schalley, C.A., Castellano, R.K., Brody, M.S., Rudkevich, D.M., Siuzdak, G. and Rebek, J. Jr (1999) J. Am. Chem. Soc., 121, 4568–4579. 37 (a) Vysotsky, M.O., Pop, A., Broda, F., Thondorf, I. and Böhmer, V. (2001) Chem. Eur. J., 7, 4403–4410; (b) Frish, L., Vysotsky, M.O., Böhmer, V. and Cohen, Y. (2003) Org. Biomol. Chem., 1, 2011–2014. 38 Thondorf, I., Broda, F., Rissanen, K., Vysotsky, M. and Böhmer, V. (2002) J. Chem. Soc. Perkin Trans., 2, 1796–1800.

39 Mogck, O., Böhmer, V. and Vogt, W. (1996) Tetrahedron, 52, 8489–8496. 40 Vysotsky, M.O., Mogck, O., Rudzevich, Y., Shivanyuk, A., Böhmer, V., Brody, M. S., Cho, Y.L., Rudkevich, D.M. and Rebek, J. Jr (2004) J. Org. Chem., 69, 6115–6120. 41 For self-assembled monolayers of heterodimers with 3, see: Xu, S., Podoprygorina, G., Böhmer, V., Ding, Z., Rooney, D., Rangan, C. and Mittler, S. (2007) Org. Biomol. Chem., 5, 558–568. 42 Castellano, R.K. and Rebek, J. Jr (1998) J. Am. Chem. Soc., 120, 3657–3663. 43 Castellano, R.K., Nuckolls, C. and Rebek, J. Jr (1999) J. Am. Chem. Soc., 121, 11156–11163. 44 Pop, A., Vysotsky, M.O., Saadioui, M. and Böhmer, V. (2003) Chem. Commun., 1124–1125. 45 (a) Warmuth, R. and Yoon, J. (2001) Acc. Chem. Res., 34, 95–105; (b) Rebek, J. Jr (2005) Angew. Chem., 117, 2104–2115; Angew. Chem. Int. Ed., 44, 2068–2078; (c) Dalgarno, S.J., Bassil, D.B., Tucker, S. A. and Atwood, J.L. (2006) Angew. Chem., 118, 7177–7180; Angew. Chem. Int. Ed., 45, 7019–7022; (d) Ajami, D., Schramm, M.P., Volonterio, A. and Rebek, J. Jr (2007) Angew. Chem., 119, 246–248; Angew. Chem. Int. Ed., 46, 242–244. 46 (a) Wendland, M.S. and Zimmerman, S. C. (1999) J. Am. Chem. Soc., 121, 1389– 1390; (b) Zimmerman, S.C., Wendland, M.S., Rakow, N.A., Zharov, I. and Suslick, K.S. (2002) Nature, 418, 399– 403; (c) Beil, J.B., Lemcoff, N.G. and Zimmerman, S.C. (2004) J. Am. Chem. Soc., 126, 13576–13577. 47 (a) Clark, T.D., Kobayashi, K. and Ghadiri, M.R. (1999) Chem. Eur. J., 5, 782–792; (b) Prins, L.J., Jolliffe, K.A., Hulst, R., Timmerman, P. and Reinhoudt, D.N. (2000) J. Am. Chem. Soc., 122, 3617–3627; (c) Jin, W., Fukushima, T., Kosaka, A., Niki, M., Ishii, N. and Aida, T. (2005) J. Am. Chem. Soc., 127, 8284–8285.

References

48 For examples involving calixarene derivatives, see: Pitarch, M., McKee, V., Nieuwenhuyzen, M. and McKervey, M. A. (1998) J. Org. Chem., 63, 946–951. 49 Vysotsky, M.O., Bolte, M., Thondorf, I. and Böhmer, V. (2003) Chem. Eur. J., 9, 3375–3382. 50 Bogdan, A., Vysotsky, M.O., Ikai, T., Okamoto, Y. and Böhmer, V. (2004) Chem. Eur. J., 10, 3324–3330. 51 Vysotsky, M.O., Bogdan, A., Wang, L. and Böhmer, V. (2004) Chem. Commun., 1268–1269. 52 Molokanova, O., Bogdan, A., Vysotsky, M.O., Bolte, M., Ikai, T., Okamoto, Y. and Böhmer, V. (2007) Chem. Eur. J. 13, 6157–6170. 53 Cao, Y., Wang, L., Bolte, M., Vysotsky, M. O. and Böhmer, V. (2005) Chem. Commun., 3132–3134. 54 Cao, Y., Rudzevich, Y., Rudzevich, V. and Böhmer, V. unpublished. 55 Wang, L., Vysotsky, M.O., Bogdan, A., Bolte, M. and Böhmer, V., (2004) Science, 304, 1312–1314. 56 Bogdan, A. and Böhmer, V. unpublished. 57 Podoprygorina, G. and Böhmer, V. unpublished. 58 Gaeta, C., Vysotsky, M.O., Bogdan, A. and Böhmer, V. (2005) J. Am. Chem. Soc., 127, 13136–13137. 59 Vysotsky, M.O. and Böhmer, V. unpublished. 60 Molokanova, O., Vysotsky, M.O., Cao, Y., Thondorf, I. and Böhmer, V. (2006) Angew. Chem., 118, 8220–8224; Angew. Chem. Int. Ed., 45, 8051–8055. 61 Thondorf, I., Rudzevich, Y., Rudzevich, V. and Böhmer, V. (2007) Org. Biomol. Chem. 5, 2775–2782. 62 Compare also: Rudzevich, Y., Rudzevich, V., Schollmeyer, D., Thondorf, I. and Böhmer, V. (2006) Org. Biomol. Chem., 4, 3938–3944. 63 (a) Castellano, R.K., Rudkevich, D.M. and Rebek, J. Jr (1997) Proc. Natl. Acad. Sci. U.S.A., 94, 7132–7137; (b) Castellano, R.K., Clark, R., Craig, S.L., Nuckolls, C. and Rebek, J. Jr (2000)

64

65

66

67

68

69

70

71

72

73

Proc. Natl. Acad. Sci. U.S.A., 97, 12418– 12421. (a) Xu, H., Stampp, S.P. and Rudkevich, D.M. (2003) Org. Lett., 5, 4583–4586; (b) Xu, H., Hampe, E.M. and Rudkevich, D.M. (2003) Chem. Commun., 2828–2829. For an intramolecular” dimerization of a molecule with a ﬂexible hexamethylene spacer see: Brody, M.S., Schalley, C.A., Rudkevich, D.M. and Rebek, J. Jr (1999) Angew. Chem., 111, 1738–1742; Angew. Chem. Int. Ed., 38, 1640–1644. For self-assembled dendrimers which are uniform in size but not necessarily in structure see: Franz, A., Bauer, W. and Hirsch, A. (2005) Angew. Chem., 117, 1588–1592; Angew. Chem. Int. Ed., 44, 1564–1567. Rudzevich, Y., Rudzevich, V., Schollmeyer, D., Thondorf, I. and Böhmer, V. (2005) Org. Lett., 7, 613–616 See also Ref. [68]. Rudzevich, Y., Vysotsky, M.O., Böhmer, V., Brody, M.S., Rebek, J., Jr, Broda, F. and Thondorf, I. (2004) Org. Biomol. Chem., 2, 3080–3084. Rudzevich, Y., Rudzevich, V., Moon, C., Schnell, I., Fischer, K. and Böhmer, V. (2005) J. Am. Chem. Soc., 127, 14168– 14169. Kanamathareddy, S. and Gutsche, C.D. (1993) J. Am. Chem. Soc., 115, 6572– 6579. A self-threaded rotaxane” was also observed as a (by) product, when two homooxacalix[3]arenes were covalently connected to a molecular capsule via three p-xylylenediamide linkers: Zhong, Z., Ikeda, A. and Shinkai, S. (1999) J. Am. Chem. Soc., 121, 11906–11907. Arduini, A., Ferdani, R., Pochini, A., Secchi, A. and Ugozzoli, F. (2000) Angew. Chem., 112, 3595–3597; Angew. Chem. Int. Ed., 39, 3453–3456. Credi, A., Dumas, S., Silvi, S., Venturi, M., Arduini, A., Pochini, A. and Secchi, A. (2004) J. Org. Chem., 69, 5881–5887.

j183

184

j 5 Tetra-urea Calix[4]arenes — From Dimeric Capsules to Novel Catenanes and Rotaxanes 74 Arduini, A., Calzavacca, F., Pochini, A. and Secchi, A. (2003) Chem. Eur. J., 9, 793–799. 75 Arduini, A., Ciesa, F., Fragassi, M., Pochini, A. and Secchi, A. (2005) Angew. Chem., 117, 282–285; Angew. Chem. Int. Ed., 44, 278–281. 76 Fischer, C., Nieger, M., Mogck, O., Böhmer, V., Ungaro, R. and Vögtle, F. (1998) Eur. J. Org. Chem., 155–161. 77 Smukste, I., House, B.E. and Smithrud, D.B. (2003) J. Org. Chem., 68, 2559–2571. 78 Li, Z.-T., Ji, G.-Z., Zhao, C.-X., Yuan, S.D., Ding, H., Huang, C., Du, A.-L.

and Wei, M. (1999) J. Org. Chem., 64, 3572–3584. 79 For an analogous example including a rigid, chiral calix[4]arene like building block see: Okada, Y., Miao, Z., Akiba, M. and Nishimura, J. (2006) Tetrahedron Lett., 47, 2699– 2702. 80 Lu, L.-G., Li, G.-K., Peng, X.-X., Chen, C.-F. and Huang, Z.-T. (2006) Tetrahedron Lett., 47, 6021–6025. 81 Böhmer, V. (1995) Angew. Chem., 107, 785–818; Angew. Chem. Int. Ed. Engl., 34, 713–745.

j185

6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding Amber L. Sadowy and Rik R. Tykwinski

6.1 Introduction

Historically, one might argue that the synthesis of many shape-persistent macrocycles (SPMs) was initiated through a combination of their enticing symmetrical structures and the synthetic challenge they posed. More recently, however, SPMs have found an important role as functional components in supramolecular chemistry. In comparison to ﬂexible macrocycles, SPMs have been designed to exhibit fewer degrees of conformational ﬂexibility, which greatly facilitates their assembly into highly organized supramolecular systems. From the onset, acetylenic building blocks have played a deﬁning role in the construction of SPMs, and, most recently, the evolution of modern metal-catalyzed coupling reactions to form sp-sp2 and sp-sp carbon bonds has greatly advanced this ﬁeld. Furthermore, SPMs based on acetylenic carbon scaffolding are particularly well suited to applications in supramolecular chemistry, given their synthetic accessibility, rigid or semi-rigid conﬁguration, and lack of steric impedance about the acetylenic units. The synthesis of acetylenic SPMs has been reviewed on a number of occasions, typically with a focus on a particular structural motif, and those molecules based on an arylene ethynylene framework have received the lions share of the attention [1–13]. Likewise, the use of acetylenic SPMs in functional systems has also been reviewed, but, again, typically with respect to a particular molecular design [5,14–16]. In view of previous literature in this area, the goal of this particular chapter will be to integrate the two aspects. First, a general summary of synthetic strategies commonly used for the synthesis of acetylenic SPMs will be described. Second, the implementation of SPMs as building blocks in supramolecular chemistry will be presented, with an emphasis on the synthetic aspects as possible. The function of SPMs has been broadly divided into three sections: (a) SPMs used as components of supramolecular assemblies, (b) SPMs used in host–guest systems, and (c) aggregation/adsorption of SPMs, including solution state aggregation, liquid crystallinity, and surface adsorption. Clearly, the substantial breadth of research that has been conducted in the area of acetylenic SPMs prevents a comprehensive description of this in the conﬁnes of a Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

186

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding single chapter. Thus, examples have been chosen to highlight the diversity of recent synthetic achievements, as well as the application of SPMs in particular areas of supramolecular chemistry. 6.1.1 SPM Synthesis through Intermolecular Reactions

One of the most straightforward routes to the synthesis of acetylenic SPMs is the one-step nA ! An approach, in which a suitable precursor is subjected to conditions that effect cyclooligomerization (Figure 6.1A). While often biased to selectively produce a desired cyclic oligomer, other macrocycles are often formed, which can lead to a tedious separation step. As with many macrocyclizations, the success of this strategy beneﬁts from highly dilute conditions that favor the kinetic formation of cyclic products (ﬁrst-order ring-closing reaction) over acyclic oligomers (secondorder reaction). To circumvent the need for extraordinary amounts of solvent to achieve high dilution, many reactions are done under pseudo-high dilution, where one or more of the reactants is added slowly to the reaction mixture. The nA ! An approach is often used for the synthesis of butadiynyl-based macrocycles, in which oxidative acetylenic homocoupling can be relied on for efﬁcient bond formation. This is demonstrated in the synthesis of expanded radialenes (1) to which mesogenic side groups have been appended, as reported by Diederich and coworkers (Scheme 6.1A) [17]. In this case, trimer, tetramer, and pentamer macrocycles were observed with tetramer as the dominant product. There are a number of variants of oxidative acetylenic homocoupling from which to choose, all based on adaptation of Glasers original report [18,19]. Hay (e.g., CuCl, TMEDA, O2, CH2Cl2) [20], Eglinton (CuSO4 or [Cu(OAc)2], py) [21,22], and Breslow (CuCl, CuCl2, oxygen-free pyridine) [23] conditions are the most commonly encountered. More recently, the use of Pd(II) as a co-catalyst in the presence of Cu(I) and O2 (or an external oxidant such as chloroacetone or I2) has emerged as an alternative to the traditional Cu(I)/Cu(II) homocoupling-based methods [24], as shown, for example, in the synthesis of TTF-based [18]annulenes 2 and 3 by Iyoda and coworkers (Scheme 6.1B) [25]. There seems, unfortunately, to be no real consensus as to how to predict which homocoupling protocol stands the best chance of success for a particular target, although it has been shown that reagent solubility, for example, can dictate the success of one method over another [26]. The most recent method developed for the nA ! An approach relies on dynamic covalent bond formation using a metathesis reaction. In this case, reactions are typically under thermodynamic control, providing the potential for increased selectivity in product formation. The initial examples using alkyne metathesis toward the formation of SPMs were reported by Adams, Bunz, and coworkers using the precatalyst [Mo(CO)6] [27,28], but rather low yields of the desired products (4) limited general applicability (Scheme 6.2). Recent efforts by Moore and coworkers using a Mo(VI)-alkylidyne catalyst, however, have reﬁned this process such that precipitation-driven reactions now provide moderate to excellent results (see Scheme 6.24) [29].

6.1 Introduction

Figure 6.1 Schematic description of synthetic strategies for macrocycle formation.

j187

188

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding Ar

A)

Ar

Ar

Ar a then b Ar

Ar

Sii-Pr3

i-Pr3Si

n

Ar = –PhO(CH2)11 O

Ar

CN

Ar

1 n = 1 4%, n = 2 56%, n = 3 4%

B)

RO2C

RO2C

CO2R

S

S

S

S

CO2R

S

S

S

S

c

S

S S RO2C RO2C

S

S

S

CO2R

S

S 2 R = Bu 29% 3 R = Oct 33%

CO2R

Scheme 6.1 (A) Synthesis of radialenes 1 by oxidative acetylenic homocoupling [17], and (B) Synthesis of TTF based [18]annulenes 2 and 3 by Pd-mediated homocoupling [25]. Reagents and conditions: (a) TBAF, THF; (b) [Cu(OAc)2], O2, py/PhH; (c) CuI, [PdCl2(PPh3)2], Et3N.

The nA ! An approach to macrocycle assembly can also derive from an unsymmetrical, bifunctional precursor (Figure 6.1B) and often proceeds using a metalcatalyzed cross-coupling reaction. This approach was used by Eglinton and coworkers to synthesize the trimeric SPM 5 via a Castro–Stephens coupling starting with the preformed Cu-acetylide 6 (Scheme 6.3) [30]. The same product was later synthesized in improved yield by Linstrumelle and coworkers using a Pd-catalyzed Sonogashira coupling, a sequence starting with the in situ removal of acetone from 7 under basic conditions to provide the terminal alkyne [31]. A variation of the one-pot approach that has, perhaps, been underdeveloped to date involves the use of multiple components (A and B), where the geometry of the reactants is designed to provide selectivity toward one macrocyclic product, nA þ nB ! An (Figure 6.1C). For example, Iyoda and coworkers exploited this approach to form selectively the cyclic trimer 8 based on the combination of a

6.1 Introduction R1

R1

R2

R1 [Mo(CO)6]

R2

R2

R1

j189

R1

R2

phenol cocatalyst

R2

Me

Me

R2

R2

R1 = t-Bu, Hex, dodecyl R2 = H, Me R1

R1

4 0.5–6%

Scheme 6.2 SPM formation by alkyne metathesis [27].

Cu I

a

b

26%

36%

Br

7

5

6

OH

Scheme 6.3 Formation of SPM 5 from unsymmetrical precursors 6 or 7. Reagents and conditions: (a) py, heat [30]; (b) NaOH, [Pd(PPh3)4], CuI, PhH, BTAC, 85 ˚C [31].

linear acetylene linker and 3,4-diiodothiophene corner piece 9 (Scheme 6.4) [32]. Considering that six new carbon-carbon bonds are in formed in this reaction, the overall yield represents an impressive 85% yield for each cross-coupling event. A major drawback of the one-pot, one-step approach is the often encountered lack of selectivity toward forming a single cyclic oligomer, coupled with a potential lack of control over which oligomer is formed preferentially. To help to overcome this difﬁculty, oligomeric fragments of the desired target are often assembled (Figures 6.1D and 6.1E) and then combined together in the ﬁnal step, helping to maximize product formation (e.g., 2(A–A) ! A4 or A–A þ A0–A0 ! A4). Once again, the ring forming-step may be a homocoupling event, as in Figure 6.1D (typically via S S

I

9

H I

H (xs)

[PdCl2(PPh3)2], CuI Et3N/DMF S

S 8 37%

Scheme 6.4 Formation of SPM 9 via a one-pot, two-component reaction [32].

190

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding N

N

A) N

N

[Cu2(OAc)4] py, rt

N

B)

F

F F

F

Hex

11

Br

F

F

F

F

Hex

+

Br

F F

F

F

Hex

Hex

[Pd2(dba)3•CHCl3]

F

F

F F

N

10 71%

i-Pr2NEt, CuI PhMe, rt

F

F

F

F F

F F

F

Hex 12

Hex 13 4%

Scheme 6.5 Intermolecular cyclization of preformed oligomeric segments by (A) Eglinton homocoupling to form 10 [33] and (B) Cadiot—Chodkiewicz heterocoupling to form 13 [34].

oxidative acetylenic homocoupling) or the combination of two complementary pieces, as in Figure 6.1E (e.g., via a Sonogashira, Cadiot–Chodkiewicz, or metathesis reaction). Baxters formation of the pyridine-containing macrocycle 10 using the Eglinton reaction provides an example of the former (Scheme 6.5A: see also Figure 6.3) [33], while Mayor and Shus use of a Pd-catalyzed Cadiot–Chodkiewicz reaction to combine the two halves (11 and 12) of macrocycle 13 demonstrates the latter (Scheme 6.5B, see also Scheme 6.23) [34]. The synthesis of these two SPMs provides a salient example of how synthetic yields in macrocycle formation can vary widely, even with the use of a similar strategy. 6.1.2 SPM Synthesis through Intramolecular Reactions

A further improvement on the concept of using preassembled oligomers exploits the involvement of a template (covalent or noncovalent) to entropically trap the pieces

6.1 Introduction

R

R

R

R

N N

N

N N

R

Zn

N N R

N

R

Zn N

j191

CuCl, TMEDA CH2Cl2, air

R R

15

R

N N

+

N

Zn

N N

R N

N

R

14 R = CH2CH2CO2CH3 70%

Scheme 6.6 Templated, selective formation of porphyrin dimer 14 [41].

in a synthetically favorable position (Figure 6.1F) [35–40]. Templation then nominally results in an intramolecular ring-closure reaction (albeit in the case where weaker non-covalent interactions template the fragments together, the deﬁnition of intramolecular may not rigorously apply). The concept of templated formation of shape-persistent macrocycles was originally introduced by Sanders and coworkers toward synthesis of porphyrin dimer 14 shown in Scheme 6.6, templated by 4,40 bipyridine [40–43]. In the absence of the 4,40 -bipyridine template, cyclization of 15 under analogous conditions led to a mixture of di- and trimeric macrocycles in reduced yield. In addition, catenation (see Schemes 6.15 and 6.16), metal templation with Pt and Cu (see Schemes 6.16 and 6.20), quadrupolar interactions of aryl and perﬂuoroaryl groups (see Schemes 6.22 and 6.23) and covalent tethering (see Scheme 6.27) have also been successfully used in the context of templated macrocycle formation. To achieve more precise control over a cyclization process, it is often desirable to reduce the cyclization to a single, intramolecular bond-forming reaction, as shown in Figure 6.1G. This approach necessitates the stepwise assembly of an acyclic oligomer lacking only this crucial ring-forming bond. While the synthesis of the requisite precursor can be tedious and time consuming relative to a one-pot cyclooligomerization, the intramolecular nature of the cyclization event ensures the highest selectivity toward formation of a single cyclic product. The route can also result in an increased product yield, although yields can remain frustratingly low even with this method. With respect to acetylenic macrocycles, two common approaches will be used to demonstrate. The ﬁrst involves the use of triazenes as a masking group for aryl iodides, a transformation introduced by Moore and coworkers in 1991 [44]. As reported by Yoshida and coworkers, the two halves 16 and 17 were assembled separately and then connected through a Sonogashira reaction to give acyclic octamer 18 (Scheme 6.7) [45]. Three subsequent steps then gave the SPM 19: (1) reaction of the aryltriazene 18 with I2 to give the aryl iodide (MeI can also be used [44]), (2) removal of the trimethylsilyl group to provide the terminal alkyne, and (3) ring closure under Sonogashira conditions.

192

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

R

R N

N

N

R

a

+ N

R

I SiMe3

N N 17

16

Et2N

R N

R

R

R

N

N b–d

N

N

N Et2N R

N

18 62%

N SiMe3

N

R N

R R = CO2Me

R

19 35% (over three steps)

Scheme 6.7 Stepwise formation of SPM 19 using an aryltriazene as a masking group for an aryl iodide [45]. Reagents and conditions: (a) [PdCl2(PPh3)2], CuI, Et3N/THF (1:1); (b) I2, ClCH2CH2Cl, 80 C; (c) KOH, MeOH, CHCl3; (d) [PdCl2(PPh3)2], CuI, Et3N/THF (1:1).

A second common technique relies on the use of orthogonal alkyne protecting groups, based on the ability to selectively remove a trimethylsilyl (or triethylsilyl) group in the presence of a triisopropylsilyl group (TMS, TES, and TIPS, respectively). The synthesis of tetramer 20 by Tobe and coworkers demonstrates this approach (Scheme 6.8) [46]. Starting with 21, selective removal of the TMS group with K2CO3/MeOH generated the terminal alkyne, which was used without isolation in a Sonogashira reaction with 22, and this was followed by a second Sonogashira reaction with TMS-acetylene to give 23. Removal of the TMS group and oxidative homocoupling gave the macrocycle precursor 24. Finally, the TIPS groups were removed with TBAF, and a homocoupling reaction gave the desired product 20. While providing low yield (10%), the preassembly of 24 helped to ensure preferential formation of the desired cyclic tetramer. In a comparison of Cu- versus Pd-homocoupling reactions, Haley and coworkers demonstrated a connection between the selectivity of the ring forming event, the

6.1 Introduction R

R SiMe3

R

R

I 22

21

j193

Br

a, b

23 92% (from 21)

i-Pr3Si

i-Pr3Si

SiMe3

c, d

R

R

R

R

e, d

R

R R

R R = CH2OC8H17

i-Pr3Si

Sii-Pr3

20 10% (from 24)

24 82% (from 23)

Scheme 6.8 Intramolecular ring closure to provide macrocycle 20 [46]. Reagents and conditions: (a) K2CO3, THF, MeOH, rt, then [Pd2(dba)3 CHCl3], CuI, PPh3, Et3N, 22, 50 C; (b) [Pd2(dba)3 CHCl3], PPh3, CuI, Et3N, TMSA, 75 C; (c) K2CO3, THF, MeOH, rt; (d) [Cu(OAc)2], py, rt; (e) TBAF, THF, rt.

catalyst used, and the ring strain present in the transition state of the bond-forming step [47,48]. The selectivity is demonstrated in the formation of bis[15]- and bis[14] annulenes (25,26 and 27,28) shown in Scheme 6.9, several derivatives of which were explored for self-association (see Figure 6.4). Brieﬂy, it was shown that intramolecular homocoupling of 29 or 30 under Eglinton conditions gave preferentially the bis [15]annulenes 25 and 26. This selectivity was ascribed to the nature of the dimeric Cu-acetylide intermediate, which, prior to reductive elimination, is commonly assumed to adopt a pseudo-trans-conﬁguration (Scheme 6.9, inset a). Molecular modeling showed that such an intermediate geometrically favors the formation of the less strained [15]annulenes over the more strained [14]annulenes. Conversely, use of Pd-catalyzed homocoupling conditions with 29 or 30 results in selective formation of the bis[14]annulenes 27 and 28. In this case, the cis-conﬁguration of the Pd-diacetylide intermediate, prior to reductive elimination, is presumably enforced by the chelating bidentate dppe ligand (Scheme 6.9, inset b), and this factor ultimately leads exclusively to formation of the more strained products 27 and 28. Finally, it is worth noting that the formation of acetylenic SPMs on solid support has also been attempted, although to date success has been limited [49]. Recent advances in the formation of acyclic oligomers by Moore and coworkers, however,

194

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

Scheme 6.9 Selective homocoupling based on Cu- or Pd-catalysis, by Haley and coworkers, in the formation of [15]annulenes 25 and 26 and [14]annulenes 27 and 28 [47]. Reagents and conditions: (a) [Cu(OAc)2], py, 60 C; (b) [PdCl2(dppe)], CuI, I2, THF, i-Pr2NH, 50 C.

suggest that this is a technique that could ultimately lend itself to selective SPM formation, as well as more facile puriﬁcation [50].

6.2 Supramolecular SPMs

The use of shape-persistent macrocycles in supramolecular chemistry has been limited only by the imagination of those working in this vast area of science. In many cases, a speciﬁc component is engineered into the SPM framework to govern directed supramolecular assembly processes, often through metal ion complexation. In other cases, the SPM has been designed with a speciﬁc size and/or conﬁguration so that it can act as a host for a guest ion or molecule. Alternatively, intermolecular interactions between the macrocycles result in aggregation that facilitates ordering of the system in the solution state, as liquid crystals, or on surfaces. Salient examples that illustrate some of these design concepts will be discussed below.

6.2 Supramolecular SPMs

j195

6.2.1 SPMs as Components in Supramolecular Assemblies

The placement of a particular functional group into the skeleton of a SPM provides rigid molecules capable of predictable self-assembly in supramolecular systems. To this end, Tykwinski and coworkers have synthesized cross-conjugated macrocycles with exocyclic pyridine donor sites to function as analogs to 4,40 -bipyridine (Scheme 6.10) [51]. For example, 3,5-diethynylpyridine was cross-coupled to vinyl triﬂate 31 using [Pd(PPh3)4], resulting in the macrocycle precursor 32. Subsequent desilylation and oxidative homocoupling under high dilution (0.5 mM) using Hay conditions in CH2Cl2 gave SPM 33 in 52% yield. The ability of 33 to function as a component in self-assembly reactions was established through reaction with the Ru-porphyrin 34 carried out in CH2Cl2 at room temperature and leading to the desired metal complex 35 (Scheme 6.11) [51]1). While 1 H NMR spectroscopy showed quantitative formation of complex 35, the isolated yield based on precipitation was only 60%, suggesting weak association between the SPM and the Ru-porphyrins. X-ray crystallographic analysis of 35 subsequently showed that unfavorable steric interactions between the phenyl rings of 35 and the pendent porphyrins were the likely origin of this weak association. The metal-directed, self-assembly of 33 with [Pt(OTf)2(PEt3)2] gave dimeric assembly 36 [52]. X-ray crystallographic analysis of 36 showed that the macrocyclic ligand 33 had assumed a boat conformation through complexation to Pt (Scheme 6.11, inset). This conformation results in the solid-state formation of bidirectional channels, that incorporate eight molecules of the crystallization solvent ClCH2CH2Cl per unit cell.

N

N

N Ph

Ph

Ph

Ph Ph

a

SiEt3

Ph

Ph

Ph

Ph

b then c

SiEt3 Ph

Ph

SiEt3 32 61%

Ph OTf 31

Scheme 6.10 Synthesis of macrocyclic ligand 33 [51]. Reagents and conditions: (a) [Pd(PPh3)4], CuI, Et2NH, THF, 55 C; (b) TBAF, THF, rt; (c) CuI, TMEDA, O2, CH2Cl2, rt.

1) A similar SPM based on isopropylidene subunits

gave the analogous complex 35 (see Ref. [51]).

Ph

Ph N 33 52%

196

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

Scheme 6.11 Self-assembly of SPM 33 using an Ru-porphyrin 34 and [Pt(OTf)2(PEt3)2], resulting in supramolecular complexes 35 [51] and 36 [52]. Inset: solid-state structure of 36 as determined by X-ray crystallography.

The possible use of this material as a porous solid was later evaluated using hyperpolarized 129 Xe NMR spectroscopy, which demonstrated the permanent porosity of the material even after the removal of the co-crystallized solvent [53]. Lees and Sun have synthesized similar macrocyclic ligands, with meta-substituted benzene ring linkers using Sonogashira and Hay reactions (Scheme 6.12) [54]. The reaction of dibromide 37 with terminal diyne 38 led solely to the formation of macrocycle 39, which was puriﬁed easily by recrystallization from CH2Cl2. In a manner similar to that described by Tykwinski (vide supra), oxidative homocoupling of 38 gave the nearly insoluble SPM 40 in only 32% yield, along with oligomeric products. The Re(I) supramolecular transition metal complexes 41 and 42 were then prepared by reﬂuxing 39 or 40 with the Re(I) diamine complex in THF. Both the SPMs (39 and 40) and their metal complexes (41 and 42) showed strong ﬂuorescence in solution. The SPMs 43 and 44 with opposing 2,20 -bipyridine donor sites for metal complexation were prepared by Schlüter and coworkers using a Sonogashira reaction to effect

6.2 Supramolecular SPMs

j197

N

37 Br

Br

N

N

a

t-Bu 38

N

N

t-Bu 39 54%

= N

b

N

OC N

c

CO

N

N

OC

( )n 40 32%

41 n = 1 62% 42 n = 2 26%

Scheme 6.12 Synthesis of SPMs 39 and 40 and formation of their metal complexes 41 and 42 [54]. Reagents and conditions: (a) [PdCl2(PPh3)2], CuI, i-Pr2NH, 70 C, 16 h; (b) CuCl, py, O2, rt, 30 h; (c) [(CH3CN)Re(CO)3(4,40 -t-Bu2bipy)](PF6), THF, 60 C.

ring closure (ring closing bonds indicated2) in Scheme 6.13A) [55,56]. While SPM 43 was formed in a reasonable yield of 28%, macrocycle 44 was inexplicably isolated in a much lower yield (14%) and could not be obtained pure. This disparity in yield for structurally similar macrocycles highlights an example of the unanticipated challenges that often face macrocycle assembly [57]. X-ray crystallographic analysis of 43 and 44 showed that the SPMs are layered in the solid state, resulting in channels that contain both co-crystallized solvent molecules and portions of the ﬂexible side chains. The bipyridine group can rotate freely, allowing for either exo- or endo-cyclic metal ion complexation. When either 43 or 44 was reacted with [RuCl2(bipy)2], however, only assemblies 45 and 46 formed, with exo-cyclic complexed ruthenium. throughout the chapter to show the ﬁnal ring closing bond.

2 PF6–

N

N Re CO

OC Re N N

2) This same convention will be used periodically

2+

( )n

c

N

N

CO

R

R

R1 =

N

N

R2

N

Sonogashira coupling here

N

R1

R1

N

47 R = OHex

N

N

N

R

43 CH2OHex, =H 44 R1 = CH2OTHP, R2 = CH2OHex

R1

R1

O

O

R

n

R2

R

[RuCl2(bipy)2]

R2

R

R

Scheme 6.13 (A) SPMs 43 and 44 with opposing bipyridine donor sites and synthesis of their metal complexes 45 and 46 [55], and (B) Polymers 47 and 48 formed via free-radical polymerization and ROMP, respectively [56].

R

B)

R2

A)

R1

N

N

N Ru(bipy)2

N

Ru(bipy)2

48 R = OHex

N

N

R2

N Ru(bipy)2

N

Ru(bipy)2

R1

R1

R

O

O

R

45 = CH2OHex, =H 55% 46 R1 = CH2OTHP, R2 = CH2OHex 80%

R1

R1

n

R2

4Cl–

4+

4 PF6–

4+ 198

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

6.2 Supramolecular SPMs

j199

This same group has reported one of the most detailed discussions of optimizing the synthesis of SPMs, using the basic macrocyclic structure found in 43 and 44. The aim was to reduce the number of steps, increase yields, and replace Stille reactions with the more efﬁcient and non-toxic Suzuki reactions [58]. The ultimate goal was to convert SPM monomers (not shown) into polymer 47 via free-radical polymerization and into polymer 48 through ring-opening metathesis polymerization (ROMP) using the Grubbs II catalyst (Scheme 6.13B). In the latter case, the protected Ru-complex of the macrocyclic monomer was required for the success of the ROMP reaction. Schmittel and coworkers envisioned the SPM 49 (containing phenanthroline units with exo-cyclic binding sites) as the key building block to form supramolecular complexes 50 and 51 (Scheme 6.14) [59]. Macrocycle 49 was obtained in 16% yield through a Sonogashira reaction [60], and simply dissolving 49 with either components 52 or 53 in the presence of [Cu(MeCN)4]PF6 in CH2Cl2 at room temperature gave quantitative self-assembly to the phenanthroline nanoboxes 50a or 50b. The authors relied on the use of steric and electronic effects from bulky aryl substituents at one end of the phenanthroline ligand 52 or 53 to thermodynamically and kinetically direct coordination toward the desired heteroleptic Cu(I) complexes [61]. Although X-ray crystallographic analysis could not be done for boxes 50a and 50b, spectroscopic analysis and MMþ calculations suggested an internal volume of >5000 Å3. The self-assembly reaction of 49 with AgPF6 gave the analogous supramolecular structures 51a and 51b. Conversion of the Ag(I) complexes 51a or 51b to the Cu(I) complexes 50a or 50b, respectively, could be accomplished through the addition of CuI to a solution of the former at room temperature. Regeneration of the Ag(I) nanoboxes 51a and 51b from 50a or 50b could then be achieved through the addition of excess AgCN. SPM 49 was also used for the formation of a supramolecular nanobasket (not shown) [62]. Macrocycles with acetylenic subunits have been used a great deal in the formation of catenanes, and much of this work is detailed in Chapter 10 by Wisner and Blight. Two pertinent examples will, however, be described here. Using the preorganization concept pioneered by Sauvage and Dietrich–Buchecker [63], Godt and Ünsal have reported the synthesis of a [2]catenane with 87-membered, non-interacting rings (Scheme 6.15) [64]. Synthesis of the [2]catenane began with the formation of 54, in which the two halves of the catenane are connected by a covalent carbonate template. Subsequent to desilylation, the terminal diyne 55 was subjected to acetylenic homocoupling under Breslow conditions and gave a mixture of the intended precatenane 56 and the non-catenated dimer (not shown) in a 2.6:1 ratio (90% combined yield). This mixture was subjected to carbonate cleavage with TBAF in THF, liberating the free [2]catenane in 63% yield (not shown). This synthesis was highly efﬁcient, puriﬁcation was simple, and large-scale reactions could be performed easily. Bäuerle and coworkers have synthesized a p-conjugated catenane, utilizing Cu(I) templation and a newly developed method for acetylenic homocoupling based on the reductive elimination of platinum to generate the desired C–C bond formation (Scheme 6.16) [65,66]. The C-shaped 57 was ﬁrst preorganized around Cu(I), resulting in the homoleptic bis-phenanthroline complex 58. After removal

Hex

N

R1 R1

R2

t-Bu 49

OMe

Sonogashira coupling here

OMe

t-Bu

+

53 R1 = Me, R2 = Br

52 R1 = H, R2 = Me

R2

R1

N

Hex

N

N

Hex

a or b

N

N

N

N

N

Hex

M N R1

N

N

OMe

Hex

OMe

Hex

Hex

MeO

Hex

MeO

N

N

N

R2

R1

R2

M

51a M = Ag(I), R1 = H, R2 = Me 100% 51b M = Ag(I), R1 = Me, R2 = Br 100%

R1

R1

M

Hex

Hex

R1

t-Bu

N

t-Bu

50a M = Cu(I), R1 = H, R2 = Me 100% 50b M = Cu(I), R1 = Me, R2 = Br 100%

t-Bu

N

R1

R2

t-Bu

R2

Hex

M

R1

R1

Scheme 6.14 Synthesis of nanoboxes 50 and 51 using metallosupramolecular assembly [59]. Reagents and conditions: (a) [Cu(MeCN)4]PF6, CH2Cl2, rt; (b) AgPF6, CH2Cl2, rt.

N

N

Hex

N

R1

N

N

N

N

4+

200

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

a

O

O

(CH2)23

54 R = SiMe3 55 R = H 77%

CO2Et

O

O

CO2Et

R

O (CH2)23

b

(CH2)23

O

Scheme 6.15 Synthesis of precatenane 56 by carbonate-templated oxidative dimerization [64]. Reagents and conditions: (a) KOH, EtOH, THF; (b) CuCl, CuCl2, py, rt.

O (CH2)23

O

(CH2)23

R

O

(CH2)23

O

56

CO2Et

O

O

CO2Et

O

(CH2)23

(CH2)23

O

6.2 Supramolecular SPMs

j201

Bu

Bu

Bu

S

S

S

S

Bu

Bu

S

S

S

Bu

Bu

Bu

S

S

57

a

Bu

S

S

Bu

S

S

N

N Cu

Bu

58 89%

S

S

Bu

S

N

S N

Bu

Bu

Bu

S

S

SiMe3 SiMe3

Bu

Bu

Bu

S

S

Bu

S

S

Bu

P

P Ph

Bu

Bu

Pt

Ph

+

Bu

Bu

SiMe3

SiMe3

b, c

Ph

Ph

Scheme 6.16 Synthesis of conjugated Cu(I) catenate 60 using metal templation and Pt reductive elimination [65]. Reagents and conditions: (a) [Cu(CH3CN)4]BF4; (b) CsF; (c) cis-[PtCl2(dppp)], rt, Et3N, PhMe, 60 h; (d) I2, THF.

Me3Si

Me3Si

Bu

N

N

S

Bu

Bu

Bu

Bu

S

Bu S

Bu

S

Bu

Bu

Bu S

S

S

S

S

S

S

Bu

Bu

Bu

Bu

S

S

S

S

Bu

Bu

S

S

59 34%

N

N Cu

Bu

60 74%

d

S N N Cu N N S

Bu

S

S N N

S

S

Bu

Bu

Bu

Bu

S

S

S

S

Bu

Bu

S

S

Bu

S

S

Bu

Bu

S

S

Bu

Bu

Bu

Ph

Pt

Ph

Ph

Ph

Bu

Bu

P

P

+

+

202

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

6.2 Supramolecular SPMs

Oct

O

H MeO

H MeO

N

N

O O MeO

N

MeO

H

trans-[PtCl2(PEt3)2]

Me O MeO

H

Et2NH, CuI CH2Cl2, rt, 12 h

N

MeO

O

O

O

Oct O H N

OMe

N

OMe

H

H

Me O

O Me

H

N

O N

O N

O

MeO

PEt3 Pt PEt3

Oct O

j203

MeO

N O

O

O

O

N

H

H

O Oct

O Oct

PEt3 Pt PEt3 61 18%

Scheme 6.17 Synthesis of rigid hydrogen-bonded SPM 61 [68].

of the four TMS groups with CsF, macrocyclization with cis-[PtCl2(dppp)] led to the catenate 59. The Pt-corners were reductively eliminated with I2 in THF, leading to catenate 60. Decomplexation with KCN gave the Cu-free p-conjugated oligothiophene-based catenane in good yield (not shown). It is worth noting that in smaller interlocked p-conjugated catenanes synthesized by this same group, Cu(I) demetalation was impossible, and therefore the Cu-free catenanes were not isolated [67]. In an interesting application of Pt-acetylide chemistry, Li and coworkers have used a preorganization concept to facilitate the self-assembly of SPM 61 (Scheme 6.17). The planar zig-zag precursor is stabilized through repeated, intramolecular three-center hydrogen bonding, and its rigid structure presents both acetylene groups aligned on the same side of the molecule. With this orientation, the formation of 61 in the presence of trans-[PtCl2(PEt3)2] and CuI should have been greatly favored, but the authors report an unexpectedly low yield of 18% for the cyclization step [68]. 6.2.2 SPMs in Host—Guest Systems

The ability to selectively functionalize SPMs, coupled with their preorganized nature, makes them quite useful as molecular hosts for a wide variety of guests. One of the ﬁrst examples capitalizing on this concept came from Diederich and coworkers, who synthesized a family of chiral binaphthyl-derived hosts including triangular, square, and rectangular derivatives (for example 62–67, Figure 6.2) via a one pot (nA ! An) Hay reaction. They subsequently studied the molecular recognition of

H

N

OMe

OMe H

O Oct

204

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding BnO BnO

BnO OBn

OBn OBn

O RO OR

O O– P O

–O

O O P O

4 Bu4N+ CO2CH3 OR OR BnO BnO

RO RO OBn OBn

n

62 n = 1, R = Bz 63 n = 2, R = Bz 64 n = 3, R = Bz

65 n = 1, R = H 66 n = 2, R = H

H3CO2C O P –O OO

O P O O– O

BnO BnO

OBn OBn 67

Figure 6.2 Examples of Diederichs saccharide receptors 62—66 [71,72] and disaccharide receptor 67 [69,70] based on R-binaphthol.

these SPMs with carbohydrates. The smaller SPM 65 was capable of selective carbohydrate recognition with several glucopyranosides, with modest enantioselectivity observed in some cases. Later generations of SPMs built on this same concept incorporated larger carbon scaffolds and modiﬁed receptor sites such as phosphate ester 67 shown below, a molecule that bonded selectively to disaccharides [69–72]. Kawase, Oda, and coworkers have synthesized a unique series of belt-shaped paraphenylacetylene macrocycles 68–71 (Scheme 6.18A) and studied their hostguest supramolecular chemistry. The precursor polyenes 72–75 were synthesized using a modiﬁed McMurry reaction, and a subsequent bromination/dehydrobromination sequence with t-BuOK gave the desired SPMs 68–71. A 4:1 mixture of 72/74 gave the corresponding mixture of SPMs 68 and 70, which could be separated by GPC, while pure samples of 73 and 75 could be obtained to provide 69 and 71 directly [73,74]. Using the same bromination/dehydrobromination procedure, SPMs 76–78 were synthesized with 1,4- or 2,6-naphthylene units, respectively, at opposing positions (Scheme 6.18B). SPMs 68 and 76 have a diameter of 1.31 nm, which is suitable for inclusion of C60, while that of 78 (1.41 nm) is nearly perfect for inclusion of C70 [75].3) The 78 C60 and 78 C70 complexes are extremely stable, with association constants Ka of 3) Association constants (Ka), were determined

by absorption spectroscopy and Stern–Völmer constants (KSV) by ﬂuorescence spectroscopy.

6.2 Supramolecular SPMs

j205

A) 1. Br2, CDCl3 2. t-BuOK, Et2O n

n

68 70 69 71

72–75 n = 4–7 B)

n=1 n=3 n=2 n=4

85% mixture 20% 4%

n

76 n = 1 77 n = 2

78

77 ⊃ 76 ⊃ C60

Scheme 6.18 (A) Synthesis of paraphenylacetylene SPMs 68—71 through bromination/dehydrobromination [74] and (B) molecular structure of SPMs 76—78 [75] and onion complex 77 76 C60 [76].

roughly 1 · 105 and 1 · 106 dm3 mol1 and Stern–Völmer constants KSV of 2.6 · 105 and 4.3 · 106 dm3 mol1, respectively. Using competitive complexation experiments, the selectivity of 78 for C70 vs. C60 was determined to be >10:1 [75]. Multicomponent onion-type complexes were then prepared, with two SPMs encompassing a single fullerene molecule, such as 77 76 C60 and 71 68 C60 (Scheme 6.18B) [76]. Initially, it was thought that charge-transfer interactions were the main driving force for fullerene-nanoring complexation. Recent studies have, however, shown that electrostatic and van der Waals energies are more signiﬁcant than charge-transfer energies, and electrostatic interactions are now considered the most likely driving force for the observed complexation [77]. The binding of metal ions is a common use of SPMs in supramolecular chemistry. Toward this goal, Baxter has synthesized multiple ortho-fused arylene-ethynylene SPMs using either Hay or Eglinton reactions (Figure 6.3). Macrocycle 79 was formed using Hay conditions, and molecular modeling studies suggested the 2,20 -bipyridine rings are in a cisoid conformation. This places all four nitrogen lone pair electrons oriented toward the interior cavity, promoting the coordination of metal ions of a comparable size. UV–vis and ﬂuorescence spectroscopy showed that 79 responded to only four of the 22 metal cations tested: Co(II), Ni(II), Cu(II) and Zn(II). Furthermore, the signal output was dependent on the metal cation, making 79 a possible multisubstrate, multiple readout ion sensor [78]. Macrocycles 80 and 81 were the ﬁrst reported 2,20 -bipyridyl twistophanes with a 6,60 -connection, and the synthetic

206

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding N

N

N N

N

N N 79

N

80 n = 1 81 n = 2

N

N

N n

N N S

S

N N

N

10

N

N

N 82

83

Figure 6.3 Baxters ortho-fused SPMs 10 and 79—83 used for metal ion binding [33,78—81].

yield of SPM 80 was marginally higher under Hay (13%) than Breslow conditions (9%). SPM 80 acted as a ﬂuorescence-quenching sensor for Cu(II) and Ag(I), while 81 functioned as a sensor only for Cu(II) [79]. Baxter synthesized SPM 10 in 71% yield using Eglinton conditions [33], and the coordination preference of 10 for metal ions was determined as Pd(II) > Co(II) Ni(II) Ag(I) > Fe(II) Hg(II). SPM 82 was synthesized in good yield (46%) using Eglinton conditions and was reported as the ﬁrst SPM combining thienyl (electron donor) and pyridyl (electron acceptor) units. Metal-binding studies showed that 82 binds preferentially to metal cations in roughly the same order as 10: Pd(II) >> Ni(II) > Co(II) > Ag(I) > Hg(II) [80]. Finally, Baxter and Dali-Youcef used the Castro–Stephens reaction for cyclotrimerization to 83, formed in only 5% yield. This unexpectedly low yield was attributed to the linear and/or cross-linked nature of the cuprate intermediate, making formation of the ﬁnal bond unfavorable. Nevertheless, macrocycle 83 was found to be a highly selective ﬂuorescence-quenching sensor for Pd(II) [81]. Using the Sonogashira reaction, Yoshida and coworkers have reported the stepwise synthesis of SPM 84, which was then converted to its butyl ester 85 (Scheme 6.19) [82]. X-ray crystallographic analysis conﬁrmed that 85 had a completely planar structure, with acetylenic bond angles of 168.7 and 171.0 . The macrocycle shows unusually intense ﬂuorescence (lem ¼ 355 nm, F ¼ 0.18, CH2Cl2), and the emission maximum shifted dramatically to 433 nm (F ¼ 0.41) upon the addition of SbCl5. Based on the 1 H NMR and UV–vis spectroscopic analyses, the luminophor was predicted to be the 1:1 complex 86. Interestingly, no other metal investigated showed any effect on the ﬂuorescence, making 85 a potentially useful ion sensor for toxic Sb(V).

6.2 Supramolecular SPMs

N

N CO2R

RO2C

SbCl5 CH2Cl2

Cl Cl Sb Cl Cl

BuO2C

R = Me

85

R = Bu

CO2Bu

N

N

84

j207

86 BuONa/BuOH

Scheme 6.19 SPMs 84—85 and a complex with Sb(V) (86) [82].

Fallis and Heuft have used a metal-templated synthesis to form the helical 1,10phenanthroline-capped metal-complexed SPMs 87 and 88 (Scheme 6.20) and their demetalated analogs (not shown). The addition of [Cu(OAc)2] (0.5 equiv) to a solution of 89 in pyridine and diethyl ether templated the formation of intermediate 90. Following the addition of excess [Cu(OAc)2] (5.5 equiv), the Cu(I)-complexed SPM 87 was isolated in an excellent yield of 84%. Treatment of 87 with aqueous KCN then provided the Cu-free SPM in 70% yield (not shown). The N,N-dibutylamine substituted analog 88 was formed in an analogous manner and taken on directly to the Cu-free SPM (39%) in a one-pot procedure. It was shown by variable temperature 13 C NMR spectroscopy that the helical Cu(I) complex 88 had a barrier to racemization of 4 kcal mol1 higher than that of the Cu-free analog [83]. MacLachlan and coworkers have formed large SPMs 91–93 using a simple, onepot, template-free, [3 þ 3] Schiff-base condensation procedure (Scheme 6.21). The yields of the macrocycles were 40–68%, and no evidence of acyclic oligomer or polymer formation was observed. This is presumably due to the reversibility of the imine condensation, which allows the most thermodynamically stable product to 2+

+

2 AcO–

R

R N

N N

N a

Cu

N b

Cu N

N N R

89

X–

N

90

Scheme 6.20 Templated, one-pot synthesis of helical 1,10-phenanthroline SPMs [83]. Reagents and conditions: (a) [Cu(OAc)2] (0.5 equiv), Et2O/py, 2 h; (b) [Cu(OAc)2] (5.5 equiv), Et2O/py.

N R

87 R = H, X = OAc 89% 88 R = NBu2, X = PF6 (not isolated)

208

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding RO

(91/94)

=

OR

H +

(92/95) H2N

OMe

O HO

NH2

R = Oct R = Dec

(93/96)

H

O

CHCl3

OH

CH3CN, ∆

MeO RO

OR

N

RO

N

N

Zn O

O

RO

Zn

O

O

N

Zn N

RO

HO

OH

O

O

N

OH HO

[Zn(OAc)2], THF

N

OR

N

N

N

OH

HO OR OR

RO

N

N

RO

94 R = Oct 80% 95 R = Dec 55% 96 R = Oct 78%

OR OR

91 R = Oct 68% 92 R = Dec 62% 93 R = Oct 40%

Scheme 6.21 Synthesis of Schiff-base SPMs 91—93 and their complexation with Zn(II) [84].

ultimately form, in this case the [3 þ 3] macrocycles. These SPMs bind to a number of metal ions, but coordination to Zn(II) proved to be the most interesting. Pristine 91–93 are weakly ﬂuorescent in THF (F ¼ 0.13–0.15), but complexation with Zn in THF at reﬂux to form 94–96 drastically increases the emission quantum yield (F ¼ 0.86–0.92). This enhanced luminescence is attributed to formation of a more rigid macrocycle after the binding of Zn(II) to the three host sites. The emission spectra of 94–96 are almost identical, and it was therefore suggested that the luminescence is localized to the site of metal complexation with little or no electronic coupling through the conjugated backbone of the molecule [84]. Strong solutionstate aggregation was later reported for 94 (as well as a branched alkyl derivative, not shown) based on intermoledular Zn. . .O interactions between the molecules [85]. 6.2.3 Aggregation and Surface Chemistry of SPMs

The organization of SPMs into well-deﬁned structures has been one of the most common goals, leading toward their application to supramolecular chemistry. The typically ﬂat conjugated skeleton of acetylenic SPMs facilitates intermolecular selfassociation, and a number of other secondary bonding interactions can be used to augment the aggregation process. Incorporation of long pendent alkyl chains about

6.2 Supramolecular SPMs

j209

the disc-like structure of SPMs can lead to columnar, discotic liquid crystals (LC), as well as the ability to form highly ordered monolayers on surfaces such as HOPG. The role of SPMs in such applications is described in this section. 6.2.3.1 Aggregation of SPMs Marsella and coworkers have reported SPM formation based on quadrupole enhanced self-assembly in solution (Scheme 6.22A) [86]. While solid-state packing based on aryl–perﬂuoroaryl interactions (ArH–ArF) is a well-established supramolecular motif [87,88], it was unique to see these electrostatic interactions manifested in a solution-state cyclization formation of 97–99 and then 100 and 101. Using Sonogashira conditions in THF and pseudo-high dilution conditions to reduce polymer formation, a mixture of terminal diynes 102 and 103 and diiodides 104 or 105 were reacted. Macrocycles 97–99 were formed in yields of 12, 10 and 30%, respectively, and the higher yield of 99 was taken as evidence of the ability of ArH–ArF electrostatic interactions to template the cyclization. The generality of this outcome was explored through the competitive coupling of diiodides 104 and 105 with terminal diyne 106 under the same conditions (Scheme 6.22B). An enhanced yield was again observed for the ArH–ArF SPM 100 (26%) in comparison to that of the ArH–ArH SPM 101 (10%). A stepwise cross coupling was proposed, in which equilibrium was established between a folded and an open intermediate. When both ArH and ArF were present, the folded intermediate facilitated cyclization. A)

R1

R1

R1

S

S [PdCl2(PPh3)2] CuI, i-Pr2NH

R1 R1 1 102 R = H

R2

R2

S

S

R1 R2

R1 R2

R2

R2

THF, 80 °C

103 R1 = F + I

R1

S

S I

S R2 R2 104 R2 = H

S 97 R1 = R2 = H

12%

98 R1 = R2 = F

10%

99 R1 = H, R2 = F

30%

105 R2 = F

B) S

104 + 105

S

S

[PdCl2(PPh3)2] CuI, i-Pr2NH

+ S

R

R

THF, 80 °C S

106

Scheme 6.22 Synthesis of SPMs 97—101 [86].

S R R 100 R = F 26% 101 R = H 10%

210

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding SPM 13 was designed by Shu and Mayor to foster intermolecular aggregation based on ArH–ArF interactions and was synthesized via two different routes [34]. The ﬁrst strategy coupled two halves of the macrocycle using a modiﬁed Cadiot–Chodkiewicz reaction, yielding 4% of macrocycle 13 after a difﬁcult puriﬁcation process (see Scheme 6.5B). The second strategy coupled dibromide 107 with diyne 108 under the same Cadiot–Chodkiewicz conditions, affording 13 in 7% yield after GPC puriﬁcation (Scheme 6.23). While the overall yield of the second strategy was slightly lower than that of the ﬁrst, product isolation in this case was much easier and more efﬁcient, making synthesis through the pentamer 107 preferred. MALDI-TOF MS of 13 identiﬁed signals consistent with aggregation, and a study of concentration dependence using 1 H NMR spectroscopy showed a consistent upﬁeld shift in the proton signals of the benzene moiety, indicative of face-to-face aggregation due to p–p stacking [89–95]. The size of these aggregates was not, however, determined. Tobe and coworkers have synthesized butadiyne-bridged SPMs 109–112 via an intramolecular Eglinton reaction to effect ring closure in yields of 27–50% (Figure 6.4) [89,93,96]. Aggregation studies between electron-poor SPMs 111–112 and electronrich SPMs 109–110 show results that are in many ways analogous to those arising from ArF–ArH quadrupole–quadrupole interactions as described above. Since this interaction requires mutual overlap of the aromatic rings, it is not unexpected to ﬁnd that SPMs with more interaction sites formed heteroaggregates more readily (i.e., 110 and 112 vs. 109 and 111). Pyridinophanes 111 and 112 also efﬁciently serve as a host for the tropylium cation, forming 1:1 and 2:1 complexes. Haley and coworkers have explored aggregation of isomeric SPMs with charge distribution biased by strong electron donors and acceptors. For example, [15]- and [14]annulenes 113–116 all form dimers in a range of solvents, but the [15]annulenes

F

F

F

Hex

F

F

F

F

F

Hex F

Hex

107

F

F

Hex

F

F

[Pd2(dba)3] i-Pr2NH, CuI PhMe, rt

F

F

F

F F

Br

Br

F

F

F

F

F Hex

F

F F Hex

F

F Hex

108

F

F

F Hex F

13 7%

Scheme 6.23 Synthesis of SPM 13 by a Pd-catalyzed Cadiot—Chodkiewicz reaction [34].

6.2 Supramolecular SPMs R

R X

R2

Bu2N

j211 R2

Bu2N

X R = CO2Oct

X

X n

R

109 110 111 112

X = CH, n = 1 X = CH, n = 3 X = N, n = 1 X = N, n = 3

R

R1

NO2

113

R1

= NBu2,

R2

R1

= NO2

114 R1 = NO2, R2 = NBu2

NO2

115

R1

= NBu2,

R2

Figure 6.4 SPMs 109—112 synthesized by Tobe and coworkers [96] and SPMs 113—116 synthesized by Haley and coworkers [48].

113 and 114 consistently show association constants about one order of magnitude higher than those of the analogous [14]annulenes 115 or 116 [48]. Tetrathiafulvalene (TTF) based [18]annulenes 117 and 118 (Figure 6.5) have been synthesized via cyclooligomerization (see Scheme 6.1B), while the [12]annulene 119 was formed under Sonogashira conditions using a terminal diyne and a diiodide (not shown) [25]. All three SPMs show strong self-aggregation in benzene and toluene, but not in CDCl3, CD2Cl2 and THF-d8. STM measurements have been carried out on samples of 117 and 118 deposited from dilute CH2Cl2 solutions onto an Au(1 1 1) surface and show that 117 and 118 have similar dimensions of about 2.2 nm long and 0.2–0.25 nm high. Helicene-based SPM (M,M,M)-120, as well as all other possible stereoisomers, have been synthesized by Yamaguchi and coworkers using a Sonogashira reaction of the corresponding diiodide and terminal diyne (Figure 6.6) [92]. Self-aggregation of these SPMs was evident in concentration dependence 1 H NMR spectroscopic studies in CDCl3 (0.1 and 10 mM). Dimer formation (even in benzene) was observed

Figure 6.5 TTF[18]annulenes 117 and 118 and TTF[12]annulene 119 synthesized by Iyoda and coworkers [25]. Inset: schematic representation of their self-assembly. Reprinted with permission from Org. Lett. 2006, 8, 1917—1920. Copyright 2006 American Chemical Society.

= NO2

116 R1 = NO2, R2 = NBu2

212

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

Figure 6.6 SPM (M,M,M)-120 and the structure of the dimer from Amber* calculations [92]. Reprinted with permission from Org. Lett. 2001, 3, 1097—1099. Copyright 2001 American Chemical Society.

by vapor pressure osmometry (VPO), but higher aggregation was not. The dimerization appeared to depend substantially on the conﬁguration of the helicene units: (M,M,M)-120 and (M,M,M)-120 > (M,P,M)-120 and (M,P,M)-120 > (M,M,M)-120 and (P,P,P)-120 > (M,P,M)-120 and (P,M,P)-120. It was later shown that an amine derivative of (M,M,M)-120 could be oxidatively dimerized using MnO2 in toluene to provide azo-linked cis- and trans-(M,M,M)/(M, M,M)-121, respectively (Figure 6.7) [97]. These isomers were easily separated by GPC and their stereochemistry was determined by UV–vis spectroscopy. Isomerization was not observed for the isomers upon heating or irradiation. VPO studies in CHCl3 indicated that trans-(M,M,M)/(M,M,M)-121 was monomeric below 0.01 mM and dimeric above 2 mM. On the other hand, cis-(M,M,M)/(M,M,M)-121 was a trimeric aggregate below 1 mM and polymeric above 1 mM. These observations were explained based on the conformation of the isomers, as depicted in Figure 6.7. Intramolecular or bimolecular aggregates are not possible for cis-(M,M,M)/(M,M, M)-121 because of the rigidity of the azo group at 120 ; therefore two cis molecules cannot exist in the same plane. These ﬁndings provided a new motif for selective self-assembly of bimolecular p-p interactions of helicenes. Höger and coworkers have recently observed that the free-base derivative of pyridinium SPM 122 (Figure 6.8) shows only weak aggregation in solution, forming tubular structures [98]. Once the pyridyl groups were alkylated to form the cationic species 122, however, well-deﬁned dimers were formed, even at very low concentrations. This behavior was explained through a conformational locking of the SPM into a boat conformation that directs the pyridinium groups toward the center of the dimer to accommodate the bromide ions. Ultimately, this hinders any further aggregation from either the top or bottom of the system.

6.2 Supramolecular SPMs

j213

Figure 6.7 Structure of cis- and trans-(M,M,M)/(M,M,M)-121 (top) and schematic description of their aggregation behavior (bottom) [97]. Reprinted with permission from J. Am. Chem. Soc. 2003, 125, 9268—9269. Copyright 2003 American Chemical Society.

Alkyne metathesis reactions have emerged as an efﬁcient and effective means of forming macrocycles, as described in detail by Prunet et al. in Chapter 2 of this book. In the case of acetylenic SPMs, Adams, Bunz, and coworkers were the ﬁrst to employ alkyne metathesis (see Scheme 6.2). Moore and coworkers have subsequently reﬁned this strategy, capitalizing on the selective formation of the thermodynamically favored product under equilibrium control using precipitation-driven alkyne metathesis [99]. These reactions were typically high yielding, and could be scaled up to provide gram quantities of the desired SPM. For example, reaction of 123 using EtCMo[NAr(t-Bu)]3 as a catalyst in the presence of p-nitrophenol gave 124 in 81% isolated yield on a milligram reaction scale and an amazing 77% yield on a gram scale (Scheme 6.24) [29,100]. SPM 124 was shown to form aggregates in solution, especially in polar solvents such as acetone [101]. An analogous metathesis reaction gave tetrameric carbazole-based SPM 125 in 84% yield as the sole product. This SPM self-assembled into nanoﬁbrils using the sol-gel method, driven by the favorable

214

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding t-Bu

CH3(CH2)15O CH3(CH2)15O CH3(CH2)15O

O(CH2)15CH3 O(CH2)15CH3 O

O N Et

O(CH2)15CH3

2 Br–

CH3(CH2)15O

Et N O

O

O(CH2)15CH3 O(CH2)15CH3 O(CH2)15CH3

CH3(CH2)15O CH3(CH2)15O

t-Bu 122

Figure 6.8 Cationic SPM 122 [98].

p-p-stacking of the planar macrocycles and hydrophobic interactions between the alkyl side chains [102]. The related SPM 126 also aggregated to give nanoﬁbrils that, when fabricated into thin ﬁlms, could be used in the detection of explosives (e.g., TNT) based on ﬂuorescence quenching [103]. Moore and Zhao have also reported the use of imine metathesis for the synthesis of SPM 127 in almost quantitative yield from 128 and 129 (Scheme 6.25) [104]. It had previously been shown that SPM 127 aggregates into columnar assemblies in acetonitrile [105], a phenomenon driven by the solvophobically favored aromatic stacking. It was thus proposed that, once 127 was formed under the metathesis reaction conditions, it was stabilized by the free energy gained from the intermolecular aggregation. This makes aggregates the thermodynamically most stable species and shifts the metathesis equilibrium in favor of cyclization. This aggregation driven formation of macrocycles presents a powerful new strategy for high-yield formation of SPMs. The same group also reported the quantitative formation of SPM 127 via a simple condensation of an amine and an aldehyde to introduce the imine bond [106]. In contrast to the work just described, in which aggregation facilitates product formation, aggregation can also lead to the demise of SPMs as a result of intermolecular close contacts that lead to polymerization. For example, Tobe and coworkers

6.2 Supramolecular SPMs

j215

R

Et≡Mo[NAr(t-Bu)]3

TgO2C

p-nitrophenol 30ºC, 22 h 123 R O R= Ph

TgO2C

CO2Tg R

R N

N

N

N

CO2Tg

TgO2C

R TgO2C

CO2Tg

124 Tg = –(CH2CH2O)3CH3 81 % (mg scale) 77 % (g scale)

R

125 R = C14H29 126 R = CO2C(CH3)2(CH2)11CH3

Scheme 6.24 Synthesis of SPM 124 by alkyne metathesis [29,100]. Inset: molecular structures of SPMs 125 and 126 [102,103].

have described the unusual spontaneous polymerization of hexameric SPM 130 at room temperature, while the tetramer (131) and octamer (132) were quite stable (Figure 6.9) [46]. Thus, in cases where the isolation of an SPM with butadiynyl groups proves challenging, it is wise to consider both whether the SPM was formed and whether it was stable. 6.2.3.2 Liquid-Crystalline SPMs The formation of liquid crystals (LC) from suitably functionalized acetylenic SPMs has often been the motivation for their synthesis. A number of groups have sought to capitalize on the planar or nearly planar disk-like geometry of SPMs to promote the formation one-dimensional columnar stacks based on a combination of p–p stacking, van der Waals, dipole, and hydrophobic interactions. For example, Heiney, Moore, and coworkers have reported the high-resolution X-ray diffraction analysis (XRD) of a tubular, discotic LC based on SPM 133, which has been synthesized via a Sonogashira reaction [106]. These studies reveal an unanticipated distortion and

216

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding CO2Tg

CO 2Tg NO 2

O2N N

N

CO2Tg

128

TgO 2C

CO 2Tg

oxalic acid

CO2Tg

N

N

MeCN

+ CO2Tg MeO N

CO2Tg

TgO 2C

OMe N

Tg = –(CH 2 CH 2 O) 3 CH3 CO2Tg

129

CO2Tg

CO2Tg 127 >99%

Scheme 6.25 Synthesis of 127 by imine metathesis [104].

doubling of an underlying hexagonal lattice with b 2a, as shown in Figure 6.10, suggesting that the unit cell must contain two crystallographically inequivalent molecules A and B. To determine the degree to which ions could be integrated into the columnar structure, SPM 133 was doped with AgOTf (1, 2, or 4 wt.%) and examined via powder XRD. At 4%, columnar order is essentially destroyed, but at 1 and 2%, order is maintained, and there is strong evidence that the silver and triﬂate ions are located in the center of the channels, offering a potential anisotropic LC ionic conductor. Tew and coworkers have reported ortho-phenylene ethynylene SPMs 134–136 with discotic LC properties (Figure 6.11) [107]. All attempts at the formation of 134–136 through the metathesis approach were discouraging; 136 could not be synthesized, while 134 and 135 were formed as mixtures. Ultimately, an intramolecular Sonogashira reaction was used. X-ray diffraction shows that symmetric 134 and 135

OctOCH2

CH2OOct

OctOCH2

n

CH2OOct

131 n = 1; 130 n = 3; 132 n = 5

Figure 6.9 Tetramer 131, hexamer 130 and octamer 132 SPMs synthesized by Tobe and coworkers [46].

6.2 Supramolecular SPMs

j217

Figure 6.10 Liquid-crystalline SPM 133 and its molecular arrangement in the unit cell [106]. Reprinted with permission from O. Y. Mindyuk, M. R. Stetzer, P. A. Heiney, J. C. Nelson, J. S. Moore, Adv. Mater. 1998, 10, 1363—1366.

self-assemble into highly ordered columnar structures, while asymmetric macrocycle 136 is less ordered. The most signiﬁcant difference between the meta-phenylene ethynylene SPMs reported by Moore (e.g., 133) and 134–136 is the size of the central cavity, which is ca. 8 Å for 133 (H to H) compared to 2.4 Å for 134 [108]. Since columnar assembly is thought to be inhibited by large internal voids, the small cavity size of 134–136 favors self-assembly. As described above, the common design of a discotic LC integrates the (semi)rigid core of an SPM with ﬂexible side chains about the periphery. An interesting twist of this concept was introduced by Höger and coworkers: Would a LC material arise from an SPM with ﬂexible side groups that ﬁll the inside of the macrocycle? SPM 137 was synthesized to probe the idea of an inverted LC, using an acetylenic homocoupling reaction to combine the two halves (Figure 6.12) [109]. X-ray crystallography shows that the alkyl chains are arranged within the macrocyclic framework, and the compound forms a mesophase at 185 C that becomes isotropic at 207 C. The long alkyl chains of 137 occupy an adaptable position on the SPM core, however, that could place them either inside or outside the cavity in the LC. It is thus unclear whether 137 represented a true inverted LC. SPMs 138–142 have subsequently been R1 134 R1 = R2 = O 135 R1 = R2 = O

O

R1 = O

O

136 R2 R1

R2 =

O

Figure 6.11 Ortho-phenylene ethynylene SPMs 134—136 synthesized by Tew and coworkers [107].

O O

O O

218

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

Figure 6.12 Macrocycles 137—142 exhibit inverted discotic liquid crystalline properties. Inset: crystal structure of 138 [110]. Rep¨ger, X. H. Cheng, A.-D. rinted with permission from S. Ho Ramminger, V. Enkelmann, A. Rapp, M. Mondeshki, I. Schnell, Angew. Chem. Int. Ed. 2005, 44, 2801—2805.

synthesized by an analogous sequence and also show thermotropic mesophases [110,111]. Given that the long alkyl chains in this case are constrained directionally into the interior of the SPM, as shown crystallographically for 138 (Figure 6.12, inset), the formation of inverted LCs is established. Preliminary guidelines toward the formation of inverted LCs have arisen from this study. First, there must be alkyl substituents that have the potential to ﬁll the internal void, and second, SPMs must not contain bulky extraannular substituents that cause the molecules to become interlocked [110]. 6.2.3.3 Adsorption of SPMs on Surfaces With the recent surge in surface analysis techniques, the characterization of surface-adsorbed SPMs has blossomed. This is particularly true for the use of scanning tunneling microscopy (STM), where the conjugated core of the SPM is expected to show a higher tunneling efﬁciency than nonconjugated portions (e.g.,

6.2 Supramolecular SPMs

j219

Scheme 6.26 Synthesis of macrocycles 143 and 144 [114]. Reagents and conditions: (a) [Pd(PPh3)4], CuI, PPh3, KOH, PhH, CH3N[(CH2)7CH3]3Cl, reﬂux. Insets: a) Schematic representation of the dense packing of ideal rhombic plates (left) and of networks of 143 (right); b) Schematic representation of the dense packing of ideal triangular plates (left) and networks of 144 (right) [115].

pendent alkyl groups), facilitating analysis [112]. Since the same intermolecular forces that provide for solution state aggregation and/or LC formation can also be exploited to organize SPMs onto surfaces, it is perhaps not surprising that structural motifs for surface-adsorbed SPMs are often similar to those described in Sections 6.2.3.1 and 6.2.3.2. Tobe and coworkers have synthesized SPMs 143 and 144 using the in situ deprotection Pd(0)-catalyzed cross-coupling protocol developed by Linstrumelle (Scheme 6.26) [31,113]. The cyclization of tetrabromotolan 145 with diethynylbenzene 146 in the presence of the phase-transfer catalyst CH3N[(CH2)7CH3]3Cl

220

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding afforded 143 in 9% yield, while the same protocol with 147 gave 144 in 5% yield [114]. It was demonstrated that 143 and 144 formed speciﬁc topologies at the liquid/ solid interface (TCB/HOPG), designated as a Kagomé lattice for 143 (Scheme 6.26, inset a) and a honeycomb structure for 144 (Scheme 6.26, inset b) [115]. The rarely observed Kagomé lattice is very interesting because of its relevance in the ﬁeld of spin-frustrated magnetic materials. Only a handful of 3D crystals have shown this lattice packing, and 143 is the ﬁrst 2D network to be reported with this motif. Whereas the rhomboid-shaped SPM 143 shows a corner-sharing hexagonal packing pattern, triangular 144 has side-sharing hexagonal packing. For both SPMs, the alkyl chains run almost perpendicular to the sides of the SPM framework, and each core thus obeys the symmetry aspects of its respective network. Höger and coworkers have used SPMs 148–151 to establish that the size of a rigid macrocycle should be balanced against the presence of ﬂexible extra- and intraannular groups that decorate the core periphery (Scheme 6.27) [116,117]. SPMs with both ester (148 and 149) and carboxylic acid (150) groups form well-ordered, structured monolayers on an HOPG surface (Scheme 6.27, inset a), whereas those containing intraannular thiol groups (151) form self-assembled monolayers on Au(1 1 1) (Scheme 6.27, inset b). The ring-closing step to 148–151 relied on acetylenic homocoupling of the two halves of the macrocycle, which typically provided only moderate yields. In an effort to increase the quantity of these macrocycles available for study, the covalent template approach has been applied, converting 152 to 153 and 154 in excellent yields over the three steps [118]. Unlike 148–151, however, attempts to melt the template-bound SPMs 153 and 154 into a LC phase were unsuccessful. The lack of LC behavior was attributed to the presence of the biphenylene bridge, which is longer than the diameter of the macrocycle, and therefore results in a non-planar conformation. Efforts then turned to the adsorption of 153 and 154 onto the surface of HOPG. STM has established that the SPMs are planarized on the HOPG surface and that the biphenylene bridge is also adsorbed onto the surface. Modeling studies have been used to show that in the absence of the solid support, the molecules prefer a nonplanar conformation and only become planar when adsorbed. STM has also been used to characterize SPM 155 (Figure 6.13A) and 155/C60, after assembling highly ordered arrays from 1,2,4-trichlorobenzene (TCB) onto a HOPG surface [119]. Monolayers of pure 155 exhibit perfect ordering over a relatively large area, but only two alkyl chains for each extraannular side group are observed. The STM image of the ordered array of 155/C60 at high resolution (Figure 6.13B–C) displays well-ordered bright spots, vastly different from the STM images of pure 155. The C60 molecules are not sequestered inside of the macrocycle, but are located at the sides of the macrocycle near the bithiophene units. This observation indicates that the driving force for the supramolecular assembly is the donor–acceptor interaction between the C60 and bithiophene groups rather than C60 and the uncovered HOPG surface inside macrocycle 155. The synthesis of SPMs 156 and 157 (Figure 6.13D) has been reported by Bäuerle and coworkers [120]. Characterization of 156 and 157 by STM at the solution/HOPG interface shows that both SPMs self-assemble to form well-ordered, stable hexagonal

6.2 Supramolecular SPMs

Scheme 6.27 Structure of SPMs 148—151 [116,117] and templated synthesis of 153 and 154 [118]. Reagents and conditions: (a) CuCl, CuCl2, py, 81%; (b) p-TsOH, MeOH, CHCl3, 99%; (c) ClCH2Ph(OR)3, K2CO3, DMF, 59% (153) and 72% (154). Inset a) STM images of 148 at the liquid/solid (1phenyloctane/HOPG) interface, 17.7 · 17.7 nm2 (left) and 12.7 · 12.7 nm2 (right); Reprinted with permission from J. Am. Chem. Soc. 2004, 126, 214—222. Copyright 2004 American

j221

Chemical Society. Inset b) STM images of SAMs of 151 on Au(1 1 1) at the air-substrate interface, 200 · 200 nm2 and z-proﬁle taken along the black line showing a row width of 5 0.4 nm (left). The same SAMs at higher magniﬁcation at the vacuum-substrate interface, 40 · 40 nm2 (right); Reprinted with permission from Langmuir 2004, 20, 2781— 2784. Copyright 2004 American Chemical Society.

222

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding

Figure 6.13 (A) SPM 155 [119]; (B) A proposed structural model for 155/C60 on HOPG; (C) A high-resolution STM image (30 · 30 nm2) of an ordered array with 155/C60 on HOPG showing the adsorption site of C60 on the macrocycle 155. Both images reprinted with permission from J. Am. Chem. Soc. 2006, 128, 4218—4219. Copyright 2006 American

Chemical Society; (D) Oligiothiophene SPMs 156 and 157 [120]; (E) Short-range ordering at solution/HOPG interface of SPM 156 from top view and 3D side view; (F) Short-range ordering at the solution/HOPG interface for SPM 157 [121]; Both images reprinted with permission from E. Mena-Osteritz, P. Ba¨uerle, Adv. Mater. 2001, 13, 243—246.

arrays [121]. The STM image of 156 shows the unique donut-like molecular shape of macrocycle 156 from a top and 3D side view (Figure 6.13E). The images agree with the semiempirical calculations that predict a spider-like conformation for the SPM, with the butyl side chains bent downwards. The STM image of 157 displays a cycle divided into two halves (Figure 6.13F). This substructure corresponds to the electron density distribution of the HOMO–1 for 157, where one nodal plane results in two C-shaped pieces of the macrocycle. The cyclic structure and highly ordered self-assembly of these molecules lend themselves to the possible formation of channels or nanotubes. Schlüter and coworkers have synthesized a number of macrocycles (158–159) with either one or two (opposing) terpyridine groups (Figure 6.14), and they provide an excellent comparison of the synthetic efﬁciency of routes used for their formation [57]. The peripheries of the macrocycles have been decorated with different ﬂexible side chains, and physical properties were assessed as a function of the nature and number of these. It is interesting to note that while all SPMs with one terpyridine

= CH2OHex, OHex

R4

= OHex, H

R1

R4 = H, Hex

R3

158

n

R4

R2 = CH2OHex, OHex, CH2OTHP

R1

n = 0, 1

n

R2

R2

R4

R4

R3

¨ter and coworkers [57]. Figure 6.14 SPMs 158 and 159 synthesized by Schlu

N

N

N

R1

N

N

N

R1

R1

n m

R3 159

R3 = H, Hex

n

R3

m

R2 = CH2OHex,CH2OTHP

R1 = CH2OHex

m = 0, 1

n = 0, 1

R3

R3

R2

R2

N

N

N

6.2 Supramolecular SPMs

j223

224

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding group (158) are nicely crystalline, those with two are not. X-ray crystallographic characterization of three derivatives provides insight into the packing effects of the various pendent alkyl chains as well as the propensity to adsorb onto the surface of HOPG. For one derivative of 158 (n ¼ 0, R1 ¼ R2 ¼ OHex, R3 ¼ H, R4 ¼ Hex), an ordered, albeit loosely packed, molecular arrangement was observed by STM on the HOPG surface. Crystallographic analysis shows that the SPM is not planar, however, a fact that likely makes surface adsorption entropically expensive, and this has thus limited the resolution of their analysis.

6.3 Conclusions

Whether by more traditional means such as acetylenic homocoupling or by more modern metathesis reactions, the synthesis of SPMs for use in supramolecular chemistry continues to evolve at an ever-increasing pace. Modern metal-mediated coupling techniques have helped to improve many ring-forming reactions, but it has been the imagination of the chemists involved that has been the most signiﬁcant advance. Once considered only as structural curiosities, acetylenic SPMs have now evolved into building blocks for an incredible range of functional materials. The synthesis of acetylenic SPMs and their applications to supramolecular chemistry will no doubt continue to challenge and fascinate the chemists who cannot resist their allure.

6.4 Experimental: Selected Procedures 6.4.1 SPM 13: Pd-Catalyzed Cadiot—Chodkiewicz Conditions

A stirred solution of [Pd2(dba)3CHCl3] (0.03 g, 0.03 mmol) and CuI (0.015 g, 0.078 mmol) in dry toluene (700 mL) and dry i-Pr2EtN (3 mL) was deoxygenated by bubbling with N2. Over a 10 h period a solution of 11 (61.4 mg, 0.10 mmol) and 12 (75 mg, 0.10 mmol) in dry toluene (120 mL) was added dropwise under an N2 atmosphere. After stirring for 4 d, the solvents were removed under reduced pressure, and the residue was dissolved in toluene (15 mL), ﬁltered, and puriﬁed by SEC and column chromatography (SiO2, toluene) to provide 13 (5.1 mg, 4.2%) as a brown solid [34]. 6.4.2 SPM 19: Use of Aryltriazene as a Masking Group for Aryl Iodides

Aryltriazene deprotection occurred in a sealable ﬂask charged with the starting material 18, I2 (1.2 equiv), and ClCH2CH2Cl. The solution was degassed, back-ﬁlled with Ar, sealed, and stirred at 80 C for 12 h. After cooling, the reaction mixture was washed three times with 10% aqueous Na2S2O3, dried over MgSO4, and

6.4 Experimental: Selected Procedures

j225

concentrated in vacuo. The residue was puriﬁed by column chromatography on SiO2 (CHCl3:EtOAc) to give the aryl iodide product. Subsequently, TMS-acetylene deprotection proceeded by dissolving the aryl iodide in an approximate 10:1 mixture of CHCl3 and MeOH with 2 M aqueous KOH (10 equiv). The reaction mixture was stirred at room temperature and monitored by TLC until reaction was complete. It was then carefully washed with water, and the organic fractions were dried over MgSO4 and evaporated. Column chromatography on SiO2 (CHCl3:EtOAc) yielded the terminal alkyne product. The cyclization step was done in dry Et3N and THF (2:1) with [PdCl2(PPh3)2] (0.1 equiv) and CuI (0.05 equiv) in a Schlenk ﬂask that had been evacuated and back-ﬁlled with Ar three times. While this mixture was stirred under Ar at 75 C, a solution of the terminal acetylene/aryl iodide in dry THF was added to the ﬂask by a syringe pump at a rate of 8 mL/h. After this addition, the reaction mixture was stirred at the same temperature for 2 h, and the solvent was then removed with a rotary evaporator. The product was puriﬁed by column chromatography on SiO2 (CHCl3:EtOAc and/or PhH:EtOAc) followed by recrystallization from CHCl3 to yield 19 (5%) [45]. 6.4.3 SPM 20: Eglinton Conditions

To a solution of [Cu(OAc)2] (546 mg, 3.00 mmol) in 200 mL of py/PhH (3:2) was added dropwise a solution of desilylated 24 (320 mg, 300 mmol) in 100 mL of the same solvent mixture over 24 h in the dark (using a Hershberg dropping funnel). After stirring at room temperature for an additional 3 h, the solvent was removed in vacuo, the green residue was passed through a short plug of SiO2, and the solvent was removed to afford crude products. The residue was puriﬁed by ﬂash chromatography followed by preparative HPLC, which afforded SPM 20 as a white solid (32.0 mg, 10%) [46]. 6.4.4 SPM 33: Hay Conditions

Compound 32 (82 mg, 0.12 mmol) was desilylated by treatment with TBAF in THF at room temperature for 30 min. Following workup, the deprotected polyyne was oxidatively coupled in the presence of CuI (0.18 g, 0.92 mmol), TMEDA (0.35 mL. 2.3 mmol), and air in dry CH2Cl2 (220 mL) for 3 h at room temperature followed by 12 h at 4 C. Following aqueous workup, the solvent was reduced by approximately 80%, resulting in the formation of a bright yellow precipitate. Isolation of the precipitate afforded 33 (29 mg, 51%) as a bright yellow solid [51]. 6.4.5 Pre-Catenane 56: Breslow Conditions

Compound 55 (1.50 g, 0.52 mmol) dissolved in py (150 mL) was added to a suspension of CuCl (6.86 g, 69.2 mmol) and CuCl2 (1.12 g, 8.30 mmol) in py (1 L) at room temperature over a period of 120 h via a syringe pump. After a stirring for an

226

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding additional 24 h, the py was distilled off, the residue was dissolved in CH2Cl2 and HCl (2N), and the organic phase was separated and washed with HCl (2N). The combined aqueous phases were extracted with CH2Cl2, the combined organic phases were then dried (MgSO4), and the solvent was removed in vacuo. Column chromatography gave a mixture of the non-catenated dimer and precatenane 56 (1.35 g, 90%, 1:2.6) as a brownish solid [63]. 6.4.6 SPM 91: Schiff-base Condensation Conditions

Under an N2 atmosphere, the corresponding dialdehyde (0.106 g, 0.40 mmol) and 1,2-dioctyloxy-4,5-diaminobenzene (0.145 g, 0.40 mmol) were combined in a solution of degassed CHCl3 (15 mL) and CH3CN (5 mL), resulting in a cloudy orange solution. This mixture was heated to reﬂux (90 C) for 24 h, giving a clear, red solution. Upon cooling, 91 precipitated as a red powder (0.16 g, 67%), which was isolated and washed with CH3CN [84]. 6.4.7 Large-scale Synthesis of SPM 124 via an Alkyne Metathesis

In an Ar-ﬁlled glove box, a solution of the molybdenum triamide catalyst (1.69 g, 2.54 mmol) and p-nitrophenol (1.06 g, 7.61 mmol) in CCl4 (200 mL) was added to a solution of the monomer precursor 125 (21.0 g, 25.4 mmol) in CCl4 (500 mL). The ﬂask was sealed and removed from the glove box. The resulting mixture was stirred for 22 h at 30 C. The precipitated by-product formed almost immediately upon warming to 30 C. After ﬁltration to remove the precipitate, the ﬁltrate was concentrated in vacuo and the residue was puriﬁed by column chromatography (i-PrOH: CH2Cl2, 1:9). The product 124 was obtained cleanly as a yellow wax (5.68 g, 77%). The major impurity was cyclic pentamer with a ratio of hexamer to pentamer of 9:1 based upon 1 H NMR integration [29].

Acknowledgments

Financial support for parts of this work, provided by the University of Alberta and the Natural Sciences and Engineering Research Council of Canada (NSERC), is gratefully acknowledged.

Symbols and Abbreviations

Ac AFM AIBN AM1

acetyl atomic force microscopy azobis(isobutyronitrile) Austin model 1

Symbols and Abbreviations

bipy Bn BTAC Bz Cy dba DME DMF dppe dppp GPC HOMO HOPG Ka KSV lem LC MALDI-TOF NBS ORTEP OTf PCBM PhH PhMe py ROMP F SEC SPM STM TBAF TCB TES THF THP TIPS TMEDA TMS TMSA tol Ts TTF VPO XRD

bipyridyl benzyl benzyltriethylammonium chloride benzoyl cyclohexyl dibenzylideneacetone 1,2-dimethoxyethane N,N0 -dimethylformamide 1,2-bis(diphenylphosphino)ethane 1,3-bis(diphenylphosphino)propane gel permeation chromatography highest occupied molecular orbital highly oriented pyrolytic graphite association constant Stern–Völmer constant wavelength of emission liquid crystal matrix-assisted laser desorption/ionization time-of-ﬂight N-bromosuccinimide Oak Ridge thermal ellipsoid plot triﬂuoromethanesulfonate [6,6]-phenyl-C61 butyric acid methyl ester benzene toluene pyridine ring-opening metathesis polymerization quantum yield size exclusion chromatography shape-persistent macrocycle scanning tunneling microscopy tetrabutylammonium ﬂuoride 1,2,4-trichlorobenzene triethylsilyl tetrahydrofuran tetrahydropyranyl triisopropylsilyl N,N,N0 ,N0 -tetramethylethylenediamine trimethylsilyl trimethylsilylacetylene tolyl tosyl (p-toluenesulfonyl) tetrathiafulvalene vapor pressure osmometry X-ray diffraction

j227

228

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding References 1 Zhang, W. and Moore, J.S. (2006) Angew. Chem. Int. Ed., 45, 4416–4439; Angew. Chem. 118, 4524–4548. 2 Zhao, D. and Moore, J.S. (2003) Chem. Commun., 807–818. 3 Young, J.K. and Moore, J.S. (1995) in Modern Acetylene Chemistry (eds. F. Diederich and P.J. Stang), WileyVCH, Weinheim, Chapter 12. 4 Höger, S. (2005) Angew. Chem. Int. Ed., 44, 3806–3808; Angew. Chem., 117, 3872–3875. 5 Höger, S. (2005) in Acetylene Chemistry (eds. F. Diederich , P.J. Stang and R.R. Tykwinski ), Wiley-VCH, Weinheim, Chapter 10. 6 Jones, C.S., O'Connor, M.J. and Haley, M.M. (2005) in Acetylene Chemistry (eds. F. Diederich, P.J. Stang and R.R. Tykwinski), Wiley-VCH, Weinheim, Chapter 8. 7 Höger, S. (1999) J. Polym. Sci., Part A: Polym. Chem., 37, 2685–2698. 8 Tahara, K. and Tobe, Y. (2006) Chem. Rev., 106, 5274–5290. 9 Spitler, E.L., Johnson, C.A. and Haley, M.M. (2006) Chem. Rev., 106, 5344– 5386. 10 Nielsen, M.B. and Diederich, F. (2005) Chem. Rev., 105, 1837–1867. 11 Marsden, J.A., Palmer, G.J. and Haley, M.M. (2003) Eur. J. Org. Chem., 2355– 2369. 12 Grave, C. and Schlüter, A.D. (2002) Eur. J. Org. Chem., 3075–3098. 13 Gleiter, R. and Hopf, H. (eds.) (2004) Modern Cyclophane Chemistry, WileyVCH, Weinheim. 14 Cheng, X.H., Ju, X.P. and Hoeger, S. (2006) Chin. J. Org. Chem., 26, 733–743. 15 Höger, S. (2004) Chem. Eur. J., 10, 1320–1329. 16 Yamaguchi, Y. and Yoshida, Z. (2003) Chem. Eur. J., 9, 5430–5440. 17 Nielsen, M.B., Schreiber, M., Baek, Y. G., Seiler, P., Lecomte, S., Boudon, C.,

18 19

20 21 22 23

24 25

26

27

28

29 30

31 32

Tykwinski, R.R., Gisselbrecht, J.-P., Gramlich, V., Skinner, P.J., Bosshard, C., Günter, P., Gross, M. and Diederich, F. (2001) Chem. Eur. J., 7, 3263–3280. Glaser, C. (1869) Ber. Dtsch. Chem. Ges., 2, 422–424. For an excellent review, see: Siemsen, P., Livingston, R.C. and Diederich, F. (2005) Angew. Chem. Int. Ed., 39, 2632– 2657; Angew. Chem., 112, 2740–2767. Hay, A.S. (1962) J. Org. Chem., 27, 3320–3321. Eglinton, G. and Galbraith, A.R. (1959) J. Chem. Soc., 889–896. Eglinton, G. and Galbraith, A.R. (1956) Chem. Ind., 737–738. O’Krongly, D., Denmeade, S.R., Chiang, M.Y. and Breslow, R. (1985) J. Am. Chem. Soc., 107, 5544–5545. Rossi, R., Carpita, A. and Bigelli, C. (1985) Tetrahedron Lett., 26, 523–526. Enozawa, H., Hasegawa, M., Takamatsu, D., Fukui, K. and Iyoda, M. (2006) Org. Lett., 8, 1917–1920. Tobe, Y., Utsumi, N., Nagano, A., Sonoda, M. and Naemura, K. (2001) Tetrahedron, 57, 8075–8083. Ge, P.-H., Fu, W., Herrmann, W.A., Herdtweck, E., Campana, C., Adams, R. D. and Bunz, U.H.F. (2000) Angew. Chem. Int. Ed., 39, 3607–3610; Angew. Chem., 112, 3753–3756. Pschirer, N.G., Fu, W., Adams, R.D. and Bunz, U.H.F. (2000) Chem. Commun., 87–88. Zhang, W. and Moore, J.S. (2004) J. Am. Chem. Soc., 126, 12796. Campbell, I.D., Eglinton, G., Henderson, W. and Raphael, R.A. (1966) J. Chem. Soc. Chem. Commun., 87–89. Huynh, C. and Linstrumelle, G. (1988) Tetrahedron, 44, 6337–6344. Iyoda, M., Vorasingha, A., Kuwatani, Y. and Yoshida, M. (1998) Tetrahedron Lett., 39, 4701–4704.

References

33 Baxter, P.N.W. (2003) Chem. Eur. J., 9, 2531–2541. 34 Shu, L. and Mayor, M. (2006) Chem. Commun., 4134–4136. 35 For an excellent recent example, see: Hoffmann, M., Wilson, C.J., Odell, B. and Anderson, H.L. (2007) Angew. Chem. Int. Ed., 46, 3122–3125; Angew. Chem., 119, 3183–3186. 36 Laughrey, Z.R. and Gibb, B.C. (2005) Top. Curr. Chem., 249, 67–125. 37 Dietrich-Buchecker, C., Colasson, B.X. and Sauvage, J.-P. (2005) Top. Curr. Chem., 249, 261–283. 38 Godt, A. (2004) Eur. J. Org. Chem., 1639–1654. 39 Anderson, S. and Anderson, H.L. (2000) in Templated Organic Synthesis (eds. F. Diederich and P.J. Stang) Wiley-VCH, Weinheim, Chapter 1. 40 Anderson, S., Anderson, H.L. and Sanders, J.K.M. (1993) Acc. Chem. Res., 26, 469–475. 41 Anderson, H.L. and Sanders, J.K.M. (1990) Angew. Chem. Int. Ed. Engl., 29, 1400–1403; Angew. Chem., 102, 1478– 1480. 42 Anderson, H.L. and Sanders, J.K.M. (1989) J. Chem. Soc. Chem. Commun., 1714–1715. 43 Walter, C.J., Anderson, H.L. and Sanders, J.K.M. (1993) J. Chem. Soc. Chem. Commun., 458–460. 44 Moore, J.S., Weinstein, E.J. and Wu, Z. (1991) Tetrahedron Lett., 32, 2465–2466. 45 Yamaguchi, Y., Kobayashi, S., Miyamura, S., Okamoto, Y., Wakamiya, T., Matsubara, Y. and Yoshida, Z. (2004) Angew. Chem. Int. Ed., 43, 366–369; Angew. Chem., 116, 370–373. 46 Nomoto, A., Sonoda, M., Yamaguchi, Y., Ichikawa, T., Hirose, K. and Tobe, Y. (2006) J. Org. Chem., 71, 401–404. 47 Marsden, J.A., Miller, J.J. and Haley, M. M. (2004) Angew. Chem. Int. Ed., 43, 1694– 1697; Angew. Chem., 116, 1726–1729. 48 Marsden, J.A., Miller, J.J., Shirtcliff, L. D. and Haley, M.M. (2005) J. Am. Chem. Soc., 127, 2464–2476.

j229

49 Shortell, D.B., Palmer, L.C. and Tour, J. M. (2001) Tetrahedron, 57, 9055–9065. 50 Elliott, E.L., Ray, C.R., Kraft, S., Atkins, J.R. and Moore, J.S. (2006) J. Org. Chem., 71, 5282–5290. 51 Campbell, K., McDonald, R. and Tykwinski, R.R. (2002) J. Org. Chem., 67, 1133–1140. 52 Campbell, K., Kuehl, C.J., Ferguson, M. J., Stang, P.J. and Tykwinski, R.R. (2002) J. Am. Chem. Soc., 124, 7266– 7267. 53 Campbell, K., Ooms, K.J., Wasylishen, R.E. and Tykwinski, R.R. (2005) Org. Lett., 7, 3397–3400. 54 Sun, S.-S. and Lees, A.J. (2001) Organometallics, 20, 2353–2358. 55 Henze, O., Lentz, D., Schäfer, A., Franke, P. and Schlüter, A.D. (2002) Chem. Eur. J., 8, 357–365. 56 Henze, O., Lentz, D. and Schlüter, A.D. (2000) Chem. Eur. J., 6, 2362–2367. 57 Grave, C., Lentz, D., Schäfer, A., Samori, P., Rabe, J.P., Franke, P. and Schlüter, A.D. (2003) J. Am. Chem. Soc., 125, 6907–6918. 58 Opris, D.M., Franke, P. and Schlüter, A. D. (2005) Eur. J. Org. Chem., 822–837. 59 Schmittel, M., Ammon, H., Kalsani, V., Wiegrefe, A. and Michel, C. (2002) Chem. Commun., 2566–2567. 60 Schmittel, M. and Ammon, H. (1999) Synlett, 750–752. 61 Schmittel, M. and Ganz, A. (1997) Chem. Commun., 999–1000. 62 Kalsani, V., Ammon, H., Jäckel, F., Rabe, J.P. and Schmittel, M. (2004) Chem. Eur. J., 10, 5481–5492. 63 Dietrich-Buchecker, C.O., Hemmert, C., Khémiss, A.-K. and Sauvage, J.-P. (1990) J. Am. Chem. Soc., 112, 8002– 8008. 64 Ünsal, Ö. and Godt, A. (1999) Chem. Eur. J., 5, 1728–1733. 65 Bäuerle, P., Ammann, M., Wilde, M., Götz, G., Mena-Osteritz, E., Rang, A. and Schalley, C.A. (2007) Angew. Chem. Int. Ed., 46, 363–368; Angew. Chem., 119, 367–372.

230

j 6 Shape-Persistent Macrocycles Based on Acetylenic Scaffolding 66 For a review of Pt-acetylide macrocycles, see: Kaiser, A. and Bäuerle, P. Top. Curr. Chem., 249, 127–201. 67 Ammann, M., Rang, A., Schalley, C.A. and Bäuerle, P. (2006) Eur. J. Org. Chem., 1940–1948. 68 Zhu, J., Wang, X.-Z., Chen, Y.-Q., Jiang, X.-K., Chen, X.-Z. and Li, Z.-T. (2004) J. Org. Chem., 69, 6221–6227. 69 Droz, A.S., Neidlein, U., Anderson, S., Seiler, P. and Diederich, F. (2001) Helv. Chim. Acta, 84, 2243–2289. 70 Droz, A.S. and Diederich, F. (2000) J. Chem. Soc. Perkin Trans., 1, 4224–4226. 71 Bähr, A., Droz, A.S., Püntener, M., Neidlein, U., Anderson, S., Seiler, P. and Diederich, F. (1998) Helv. Chim. Acta, 81, 1931–1963. 72 Anderson, S. Neidlein, U. Gramlich, V. and Diederich, F. (1995) Angew. Chem. Int. Ed. Engl., 34, 1596–1600; Angew. Chem., 107, 1722–1725. 73 Kawase, T., Darabi, H.R. and Oda, M. (1996) Angew. Chem. Int. Ed. Engl., 35, 2664–2666; Angew. Chem., 108, 2803– 2805. 74 Kawase, T., Ueda, N., Tanaka, K., Seirai, Y. and Oda, M. (2001) Tetrahedron Lett., 42, 5509–5511. 75 Kawase, T., Tanaka, K., Seirai, Y., Shiono, N. and Oda, M. (2003) Angew. Chem. Int. Ed., 42, 5597–5600; Angew. Chem., 115, 5755–5758. 76 Kawase, T., Tanaka, K., Shiono, N., Seirai, Y. and Oda, M. (2004) Angew. Chem. Int. Ed., 43, 1722–1724; Angew. Chem., 116, 1754–1756. 77 Kawase, T. and Oda, M. (2006) Pure Appl. Chem., 78, 831–839. 78 Baxter, P.N.W. (2001) J. Org. Chem., 66, 4170–4179. 79 Baxter, P.N.W. (2002) Chem. Eur. J., 8, 5250–5264. 80 Baxter, P.N.W. (2004) J. Org. Chem., 69, 1813–1821. 81 Baxter, P.N.W. and Dali-Youcef, R. (2005) J. Org. Chem., 70, 4935–4953. 82 Kobayashi, S., Yamaguchi, Y., Wakamiya, T., Matsubara, Y., Sugimoto,

83

84

85

86 87

88

89

90 91

92

93

94

95 96

97

K. and Yoshida, Z. (2003) Tetrahedron Lett., 44, 1469–1472. Heuft, M.A. and Fallis, A.G. (2002) Angew. Chem. Int. Ed., 41, 4520–4523; Angew. Chem., 114, 4702–4705. Ma, C., Lo, A., Abdolmaleki, A. and MacLachlan, M.J. (2004) Org. Lett., 6, 3841–3844. Ma, C.T.L. and MacLachlan, M.J. (2005) Angew. Chem. Int. Ed., 44, 4178–4182; Angew. Chem., 117, 4250–4254. Marsella, M.J., Wang, Z.-Q., Reid, R.J. and Yoon, K. (2001) Org. Lett., 3, 885–887. Coates, G.W., Dunn, A.R., Henling, L.M., Dougherty, D.A. and Grubbs, R.H. (1997) Angew. Chem. Int. Ed. Engl., 36, 248–251; Angew. Chem., 109, 290–293. Coates, G.W., Dunn, A.R., Henling, L. M., Ziller, J.W., Lobkovsky, E.B. and Grubbs, R.H. (1998) J. Am. Chem. Soc., 120, 3641–3649. Tobe, Y., Utsumi, N., Kawabata, K., Nagano, A., Adachi, K., Araki, S., Sonoda, M., Hirose, K. and Naemura, K. (2002) J. Am. Chem. Soc., 124, 5350– 5364. Lin, C.-H. and Tour, J. (2002) J. Org. Chem., 67, 7761–7768. Höger, S., Bonrad, K., Mourran, A., Beginn, U. and Möller, M. (2001) J. Am. Chem. Soc., 123, 5651–5659. Nakamura, K., Okubo, H. and Yamaguchi, M. (2001) Org. Lett., 3, 1097–1099. Tobe, Y., Utsumi, N., Nagano, A. and Naemura, K. (1998) Angew. Chem. Int. Ed., 37, 1285–1287; Angew. Chem., 110, 1347–1349. Shetty, A.S., Zhang, J. and Moore, J.S. (1996) J. Am. Chem. Soc., 118, 1019– 1027. Zhang, J. and Moore, J.S. (1992) J. Am. Chem. Soc., 114, 9701–9702. Tobe, Y., Nagano, A., Kawabata, K., Sonoda, M. and Naemura, K. (2000) Org. Lett., 2, 3265–3268. Saiki, Y., Sugiura, H., Nakamura, K., Yamaguchi, M., Hoshi, T. and Anzai, J.

References

98

99

100 101

102

103

104 105 106

107

108 109

110

(2003) J. Am. Chem. Soc., 125, 9268– 9269. Klyatskaya, S., Dingenouts, N., Rosenauer, C., Müller, B. and Höger, S. (2006) J. Am. Chem. Soc., 128, 3150– 3151. Zhang, W., Brombosz, S.M., Mendoza, J.L. and Moore, J.S. (2005) J. Org. Chem., 70, 10198–10201. Zhang, W. and Moore, J.S. (2005) J. Am. Chem. Soc., 127, 11863–11870. Lahiri, S., Thompson, J.L. and Moore, J. S. (2000) J. Am. Chem. Soc., 122, 11315– 11319. Balakrishnan, K., Datar, A., Zhang, W., Yang, X., Naddo, T., Huang, J., Zuo, J., Yen, M., Moore, J.S. and Zang, L. (2006) J. Am. Chem. Soc., 128, 6576– 6577. Naddo, T., Che, Y., Zhang, W., Balakrishnan, K., Yang, X., Yen, M., Zhao, J., Moore, J.S. and Zang, L. (2007) J. Am. Chem. Soc., 129, 6978– 6979. Zhao, D. and Moore, J.S. (2003) Macromolecules, 36, 2712–2720. Zhao, D. and Moore, J.S. (2002) J. Org. Chem., 67, 3548–3554. Mindyuk, O.Y., Stetzer, M.R., Heiney, P. A., Nelson, J.C. and Moore, J.S. (1998) Adv. Mater., 10, 1363–1366. Seo, S.H., Jones, T.V., Seyler, H., Peters, J.O., Kim, T.H., Chang, J.Y. and Tew, G. N. (2006) J. Am. Chem. Soc., 128, 9264– 9265. Zhang, J. and Moore, J.S. (1994) J. Am. Chem. Soc., 116, 2655–2656. Höger, S., Enkelmann, V., Bonrad, K. and Tschierske, C. (2000) Angew. Chem. Int. Ed., 39, 2268–2270; Angew. Chem., 112, 2355–2358. Höger, S., Cheng, X.H., Ramminger, A.-D., Enkelmann, V., Rapp, A., Mondeshki, M. and Schnell, I. (2005) Angew. Chem. Int. Ed., 44, 2801–2805; Angew. Chem., 117, 2862–2866.

j231

111 For a more detailed discussion of the LC and surface chemistry of SPMs such as 141–142, see: Cheng, X., Ver Heyen, A., Mamdouh, W., Uji-i, H., De Schryver, F., Höger, S. and De Feyter, S. (2007) Langmuir, 23, 1281–1286. 112 De Feyter, S., Gesquière, A., AbdelMottaleb, M.M., Grim, P.C.M., De Schryver, F.C., Meiners, C., Sieffert, M., Valiyaveettil, S. and Müllen, K. (2000) Acc. Chem. Res., 33, 520–531. 113 Ohkita, M., Ando, K., Suzuki, T. and Tsuji, T. (2000) J. Org. Chem., 65, 4385– 4390. 114 Sonoda, M., Sakai, Y., Yoshimura, T., Tobe, Y. and Kamada, K. (2004) Chem. Lett., 33, 972–973. 115 Furukawa, S., Uji-i, H., Tahara, K., Ichikawa, T., Sonoda, M., De Schryver, F.C., Tobe, Y. and De Feyter, S. (2006) J. Am. Chem. Soc., 128, 3502–3503. 116 Fischer, M., Lieser, G., Rapp, A., Schnell, I., Mamdouh, W., De Feyter, S., De Schryver, F.C. and Höger, S. (2004) J. Am. Chem. Soc., 126, 214–222. 117 Borissov, D., Ziegler, A., Höger, S. and Freyland, W. (2004) Langmuir, 20, 2781–2784. 118 Ziegler, A., Mamdouh, W., Ver Heyen, A., Surin, M., Uji-i, H., Abdel-Mottaleb, M.M.S., De Schryver, F.C., De Feyter, S., Lazzaroni, R. and Höger, S. (2005) Chem. Mater., 17, 5670–5683. 119 Pan, G.-B., Cheng, X.-H., Höger, S. and Freyland, W. (2006) J. Am. Chem. Soc., 128, 4218–4219. 120 Krömer, J., Rios-Carreras, I., Fuhrmann, G., Musch, C., Wunderlin, M., Debaerdemaeker, T., Mena-Osteritz, E. and Bäuerle, P. (2000) Angew. Chem. Int. Ed. 39, 3481– 3486; Angew. Chem., 112, 3623–3628. 121 Mena-Osteritz, E. and Bäuerle, P. (2001) Adv. Mater., 13, 243–246.

j233

7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles Laura Pirondini and Enrico Dalcanale

7.1 Introduction

Synthetic macrocycles have played a major role in shaping the ﬁeld of supramolecular chemistry. The study of their host-guest properties has been pivotal in deﬁning and understanding non-covalent interactions, which constitute the heart of supramolecular chemistry. Afterwords, the mastering of such weak interactions has allowed the emergence of self-assembly as a new tool to generate complex and otherwise elusive molecular architectures. Calixarenes [1], cavitands [2], and clefttype receptors [3] represent three-dimensional concave surfaces capable of binding complementary organic molecules and ions. Their association into larger assemblies has been pursued initially via covalent bonding, thanks to the seminal work of Cram (carcerands) [2] and Collet (cryptophanes) [4]. Their studies have triggered the interest of the chemical community in molecular containers and their peculiar inclusion properties. The drawback of the covalent approach, i.e. the need for multistep, time-consuming synthetic procedures, has been overcome by the introduction of self-assembly protocols, with H-bonding and metal coordination as the preferred options among non-covalent interactions [5]. The diversity of available objects in terms of size and shape has dramatically increased, thanks to the inherent modularity of the self-assembly approach. Metal-directed self-assembly in particular offers the following advantages [6]: (a) high directionality of the metal-ligand coordinative bonds leading to many possible coordination geometries, (b) great versatility due to diverse transition-metal complexes and to a broad spectrum of available multidentate ligands, (c) ﬁne tuning of the kinetic lability of the complexes to different self-assembly conditions, and (d) control of the polarity and charge state of the resulting objects via formation of charged or neutral complexes. The modulation of the metal-directed self-assembly protocols according to the above-mentioned guidelines introduces additional features for non-covalent molecular containers with respect to their covalent counterparts. Reversibility, selection,

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

234

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles and, above all, predictable and reproducible responses to external stimuli are the major ones. Synthetic macrocycles are ideally suited for the self-assembly of 3D objects because they offer three-dimensional scaffolds amenable to multiple ligand functionalization. Type, number, and relative orientation of the ligands can be engineered according to the ﬁnal object to be assembled. Also, the degree of rigidity of multidentate macrocyclic ligands can be ﬁnely tuned according to the requirements of the self-assembly protocol. All potential applications of molecular containers revolve around a major issue: the compartmentalization of the included guests. Drug delivery [7], catalysis [8], storage [9], separation technology [10], and stabilization of reactive intermediates [11] are typical examples in which compartmentalization determines the function. For some applications envisioned, however, compartmentalization alone is not enough. In the case of drug delivery, for instance, the possibility of releasing a guest following external stimuli constitutes an essential feature. Needless to say, this additional feature is much more easily realized using non-covalent interactions [12]. Decisive factors to be considered in designing molecular containers are: (a) size of the inner cavity, which in most cases must be of nanoscale dimensions to accommodate either single nanoscale guests or multiple smaller molecules [13], (b) size and shape of the cavity portals, to govern the communication with the exterior, and (c) solubility in an aqueous environment, to take advantage of the hydrophobic effect in driving the inclusion of non-polar, neutral guests. Another important sector where metal-directed self-assembly offers unparalleled advantages over covalent synthesis is the generation of 3D objects on surfaces. The main issue in this case is the quantitative, error-free generation of 3D architectures directly on surfaces, made possible by the thermodynamic control and reversibility of self-assembly. The transfer of single-output self-assembly protocols from solutions to surfaces provides a unique entry into highly organized, defect-free, hybrid inorganic-organic materials, which are in high demand for nanotechnology and molecular electronics [14]. In this chapter, we cover the metal-directed self-assembly of 3D architectures using macrocycles as molecular scaffolds. Particular emphasis will be given to the synthetic strategies leading to coordination cages and to multitopic macrocyclic receptors, as well as to their complexation properties.

7.2 Coordination Cages

The need for highly preorganized, multidentate ligands presenting concave surfaces has led some researchers to resort to calixarenes, resorcinarenes, and cavitands as sources of molecular scaffolds for the self-assembly of coordination cages. They have almost unlimited potential as molecular receptors and building blocks for higher architectures, and their inherent half-bowl-shaped structure is particularly advantageous in the design of molecular capsules. The topic is treated in terms of increasing complexity, expressed by the number of macrocyclic subunits connected together.

7.2 Coordination Cages

OR

R R R O O O

OR

O

6CF3SO3

2 CH2

LPd

LPd

O

N

Cu O

O

O

8+ O 8CF3SO3-

-

N N

O

O

O

O

O

6+

N PdL N

N

N

N

LPd N

N

O

N LPd

PdL

PdL N

N

O

O

N

CH2

O

OR

OR

R = CH2CH2OCH3 1

j235

2

O O O O R R R

R = CH2C(O)OCH2CH3

O O

O O

L = dppp

L = dppp

2

3

Figure 7.1 Calixarene-based dimeric capsules.

7.2.1 Dimeric Calixarene-based Coordination Cages

In the speciﬁc case of calixarenes, there are very few studies related to metal-mediated self-assembly into cage molecules. Early examples reported by Shinkai and coworkers include bis(acac)-calix[4]arene dimer (1) held together by Cu2þ cations [15] and the dimeric capsule 2 (Figure 7.1), in which two homooxacalix[3]arene units bearing 4-pyridyl groups are linked through coordination to Pd(II) metal complexes [16]. This study was particularly signiﬁcant because it showed that capsule formation requires the favorable preorganization of the building blocks: only calixarenes in a cone conformation [R ¼ CH2C(O)OC2H5] yield a dimeric molecular capsule with D3h symmetry, while conformational mobility (R ¼ CH3) switches off capsule formation. Other important prerequisites are Py-Pd(II)-Py angle close to 90 , a precise 2 : 3 calixarene/PdLn concentration ratio, and an appropriate solvent choice. Further studies on cage 2 revealed that this capsule can speciﬁcally include [60]fullerene [17] with a Kass of 39 M1 at 30 C. The kinetically stable capsule-fullerene complex was characterized by 1 H and 13 C NMR spectroscopy. Liþ-binding to the lower rims enhances the [60]fullerene inclusion ability of 2 by a factor of approximately 54, whereas Naþ-binding removes it entirely [18]. In 2001 Shinkai and coworkers reported the self-assembly of the coordination cage 3 composed of two conetetrapyridylcalix[4]arene bis-crown ligands linked through four cis-Pd(II) complex molecules [19]. The presence of the short crown linkages at the lower rim forces the calixarene into a rigid cone conformation, preventing entropically favorable intramolecular bonds. Again the rigidity and preorganization of the calixarene scaffold provides a prerequisite for capsule formation.

236

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles A different approach to the self-assembly of a supramolecular calix[4]arene capsule through metal-ligand interaction was described by Baldini et al. [20]. In this case, bisand tetra-functionalization of the calixarene upper rim with metalloporphyrins gives subunits that dimerize in a face-to-face fashion when an appropriate bidentate ligand such as DABCO is added (Figure 7.2). The originality of this approach consists in the reversal of the metal-ligand relationship usually employed in the self-assembly of coordination cages: the metal centers are introduced on the calixarene, and bidentate ligands are used as linkers. The calix-bisporphirin 4 forms the intermolecular 2 : 2 sandwich complex 6 with 1 equivalent of DABCO, generating a large cavity capable of entrapping guests of appropriate dimensions. When more than 1 equiv of DABCO is added, the system enters a fast exchange regime, leading to the formation of the open 4(DABCO)2 complex. The calix-tetraporphyrin 5, on the other hand, forms four different complexes with DABCO, depending on the stoichiometry (1 : 1; 2 : 4; 1 : 3; 1 : 4). During the addition of DABCO, all four complexes are populated, giving both intramolecular and intermolecular 5: DABCO sandwiches. The complicated equilibria between the complexes were studied using both UV-visible and 1 H NMR spectroscopy. At the precise 5:DABCO ratio of 2 : 4, the major species is the dimeric cage 7, but unfortunately it is not the only complex in solution. This example highlights the difﬁculty of driving the self-assembly equilibrium toward a single thermodynamic minimum. The presence of intramolecular complexes between porphyrin moieties on the same calix[4]arene is due to their residual conformational mobility. Cotton et al. showed the assembly of a very stable coordination cage formed by linking two calixarene scaffolds functionalized with four carboxylic acid groups via bimetal units, (DaniF)2Rh2 [21]. This capsule permanently includes an NEt4þ cation and has a BF4 counterion external to the cage, as shown by the X-ray crystal structure determination. 7.2.2 Cavitand-based Dimeric Coordination Cages 7.2.2.1 The Apical Functionalization Approach Cavitand-based coordination cages are receiving increasing attention because of the versatility of the cavitand platforms in terms of synthetic modularity and molecular recognition properties. The vast majority of coordination cages of this class are dimeric, composed of two concave macrocycles facing each other, connected through metal centers. The introduction of the ligands at the upper rim of the resorcinarene skeleton has been obtained following two different strategies: (a) functionalization at the apical positions, and (b) introduction of the ligands as bridging units between the phenolic OHs. In this section we analyze examples of such cages assembled from cavitands bearing ligand moieties at the apical positions. In this way up to four ligands can be introduced at the upper rim of a single cavitand in a diverging spatial orientation. The ﬁrst communication on cavitandbased coordination cages was reported by our group in 1997 [22]. Two tetracyano cavitands were connected through four Pd(II) or Pt(II) square-planar complexes. The self-assembly is quantitative, the cage forming immediately as the only product on simply mixing tetradentate cavitand ligands 8 with [M(dppp)(CF3SO3)2] (M ¼ Pd, Pt)

7.2 Coordination Cages

Figure 7.2 Calix-bisporphyrin and calix-tetraporphyrin and their sandwich complexes with DABCO (Ar ¼ p-pentylphenyl).

in a 1 : 2 molar ratio at room temperature (Scheme 7.1). In the presence of excess cavitand, a mixture of cage and free cavitand is obtained; on the other hand if an excess of metal precursor is added, the reaction leads to the cage plus unchanged metal precursor. Control of the self-assembly process was achieved through metalligand exchange (Scheme 7.2). Addition of eight equivalents of a competing ligand

j237

238

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Scheme 7.1 Self-assembly of Pd and Pt cage molecules 9a—k.

Scheme 7.2 Control of the self-assembly of cage 9d. (a) 8 eq. NEt3; (b) 8 eq. CF3SO3H. L ¼ dppp.

7.2 Coordination Cages

Figure 7.3 Side and top views of the molecular structure of 9d. The eight undecyl feet are omitted. Reproduced with permission from Ref. [23].

such as NEt3 to the preformed 9d led to a complete and clean disassembly of the cage into its cavitand components and [Pt(dppp)(NEt3)(CF3SO3)2]. Subsequent addition of eight equivalents of triﬂuoromethanesulfonic acid restored the original Pt complex, which immediately and quantitatively reassembled the cage. The presence of two signals in the 19 F NMR spectra in a 1 : 7 integral ratio indicates encapsulation of one of the eight triﬂuoromethanesulfonate counterions inside the cage. An X-ray crystal structure of 9d (Figure 7.3) conﬁrms the inclusion of a single, desolvated triﬂuoromethanesulfonate anion inside the cavity [23]. A more comprehensive study published in 2001 showed the self-assembly of a whole family of coordination cages 9a–k [23]. The main factors controlling cage selfassembly (CSA) have been identiﬁed as (a) a P-M-P angle close to 90 between the chelating ligand and the metal precursor, (b) Pd and Pt as metal centers, (c) a weakly coordinated counterion in the precursor complex, and (d) preorganization of the tetradentate cavitand ligand. Regarding this last point, the inﬂuence of rigidity of the tetradentate cavitand ligand has been evaluated by introducing conformationally mobile ethylene bridges instead of the more rigid methylene ones. Table 7.1 reports the results of the self-assembly process between ethylene-bridged cavitands 10 and several metal precursors compared to the case of 8. In the case of 10, CSA is possible only with Pt complexes (Scheme 7.3), which provide sufﬁcient coordinative strength toward nitriles to freeze the cavitand in the required cone conformation. Their Pd analogs fail to give CSA under the same conditions. Moreover, in all cases, the CSA is partial at 300 K and complete only above 350 K. Calorimetric measurements on the formation of cages 9, together with 1 H and 19 F NMR experiments, indicated that CSA is entropy driven, with the enthalpic cost of desolvating the molecular components, and particularly the encapsulated anion, compensated by the large entropic gain associated with the release of solvent molecules into the bulk. Desolvation is pivotal also in determining selectivity in anion encapsulation, together with cavity size. As proven in ESI-MS measurements, only in the case

j239

240

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles Table 7.1 Inﬂuence of cavitand preorganization on CSA.

Entry

Metal Precursor

Cavitand 8b

1 2 3 4 5 6 7

Pd(dppp)(CF3SO3)2 Pd(dppp)(BF4)2 Pd(dppp)(PF6)2 Pd(dppb)(CF3SO3)2 Pt(dppp)(CF3SO3)2 Pt(dppp)(BF4)2 Pt(dppp)(PF6)2

CSA CSA CSA CSA CSA CSA CSA

(9b) (9h) (9k) (9f) (9d) (9i) (9j)

Cavitand 10b No CSA No CSA No CSA No CSA CSA (11b) CSA (11c) CSA (11d)

of encapsulated BF4 is there enough room in the cavity for one molecule of solvent (CHCl3 or acetone), while in the case of CF3SO3 and PF6 the anion is included alone. This explains the discrepancy between the observed (BF4 CF3SO3 PF6) and the expected (CF3SO3 > PF6 > BF4) selectivity in anion encapsulation. The predicted selectivity trend is dictated by the relative enthalpy of desolvation of the anions, but since in the case of BF4 one molecule of solvent is included, this trend is biased toward BF4 in the observed data. Recently, dynamic and structural NMR studies have been carried out by Macchioni and coworkers in collaboration with our group [24]. The combined use of PGSE, NOE, and EXSY NMR techniques has helped determine the interionic structure, kinetic stability, and degree of anion encapsulation of coordination cages 9b,c,d in solution. Both the interionic structure and kinetic stability are solvent dependent, while anion encapsulation is not. In chloroform, all counterions are positioned around the cationic cage and move with it in solution. Addition of a more polar solvent like nitromethane disrupts the ionic integrity, leaving only the included CF3SO3 plus one or two anions

Scheme 7.3 Cage self-assembly using ethylene-bridged cavitands 10a,b.

7.2 Coordination Cages

connected to the cage. On the other hand, anion encapsulation is quantitative, regardless of the polarity of the solvent. Concerning the kinetic stability of the cage, it can be tuned by changing the coordinating metal and the solvent. The rate constants for the formation/dissociation processes of the Pd cage 9b are accelerated by more than one order of magnitude on passing from a low-polarity solvent such as chloroform-d (kobs(s1) ¼ 0.30 0.04) to a polar one, namely, nitromethane-d3 (kobs(s1) ¼ 5.2 0.8). The exchange process is remarkably slower when Pt cage 9c is considered, where the kinetically more inert Pt complex replaces the Pd one. The quest for cages with internal cavities of nanoscopic dimensions able to encapsulate large guests led our group, in collaboration with de Mendozas, to enlarge the size of the reacting ligands [25]. In particular, phenyl groups were chosen as spacers introduced in the apical position of the resorcinarene in order to extend the cavity size, while retaining the relative orientation of the cyano groups and the rigidity of the cavitand framework. Cavitands 12a,b were synthesized from tetrakis(bromophenyl)-cavitands through a reaction with CuCN. The corresponding cages 13a,b were obtained in a quantitative yield by mixing 12a,b with [Pt(dppp)(CF3SO3)2] in a 1 : 2 molar ratio at room temperature (Scheme 7.4). In this case no permanent inclusion of the CF3SO3 counterions was observed by 19 F NMR even at low temperatures: rapid exchange with the external counterions occurs because of the

Scheme 7.4 Self-assembly of enlarged Pt and Pd cage molecules 13a—c.

j241

242

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.4 Cavitand ligands with four pyridyl (14) and pyridylethynyl (15) substituents at the apical position.

large dimensions of the lateral portals. Enlarging the size of the cavitand ligand leads to the self-assembly of nanosize coordination cages; the drawback is the concomitant widening of the lateral portals, which prevents permanent inclusion of suitable guests. The self-assembly of the corresponding Pd cage 13c was attempted using [Pd(dppp)(CF3SO3)2] as a metal complex. The resultant product showed 1 H and 31 P NMR spectra consistent with the corresponding cage formation; however, no conﬁrmation came from MALDI-TOF or ESI-MS analyses. This different behavior is because the CN-Pd bond is weaker than the CN-Pt bond. A different synthetic procedure to introduce cyanophenyl groups at the apical position (12c) was reported by Kobayashi et al. in 2004, together with the synthesis of the two new tetradentate cavitand ligands 14 and 15 (Figure 7.4) [26]. A 2 : 4 mixture of 14 and [Pd(dppp)(CF3SO3)2] led to the formation of various aggregates which were not related to the expected coordination cage. This behavior is mainly due to a hindered rotation of the pyridyl groups directly attached to the apical positions. By introducing an ethynyl spacer, pyridylethynyl cavitand 15 self-assembled into the corresponding cage 16 quantitatively through a concerted process (Figure 7.5). The inability of cavitand 14 to form the corresponding homocage has been exploited in a unique and very elegant way to exclusively produce heterocages in solution. In fact, heterocage 17 was self-assembled as a sole product by mixing 12c : 14 : [Pd(dppp)(CF3SO3)2] in a 1 : 1 : 4 molar ratio. On the other hand a 1 : 1 : 4 mixture of 12c : 15 : [Pd(dppp)(CF3SO3)2] exclusively gave the two homocages 13c and 16 in a 1 : 1 ratio. The authors have therefore demonstrated that by tuning the coordination ability of the cavitand ligands on the basis of the kinetic stability of their metal complexes and their steric demand it is possible to direct the self-assembly process towards either exclusively homo-or heterocages. Further studies on the kinetic and thermodynamic stabilities of Pd cages 13c, 16, 17, and the corresponding Pt cages have been reported by Kobayashi and coworkers in a subsequent paper [27]. The stability of the cages increases in the order 13c < 17 < 16.

7.2 Coordination Cages

Figure 7.5 Homocage 16 and heterocage 17.

A different approach to cage widening has been proposed by Sherburn, Stang, and coworkers [28]. Using large dinuclear Pt precursors like [2,9-(trans-Pt(PEt3)2NO3)2phenanthrene] in connection with bis-4-pyridyl cavitands 18 and 19, they assembled expanded coordination cages 20 and 21, which have wide lateral portals (Scheme 7.5). This strategy offers a useful alternative to the often synthetically demanding introduction of long, rigid ligands on the cavitand upper rim. It also highlights one of the hallmarks of self-assembly: modularity. Other examples of coordination cages prepared by self-assembly of cavitand derivatives were shown by Hong and coworkers starting from the more ﬂexible tetradentate cavitands 22a–c (Scheme 7.6) [29]. Also in this case, cages 23a–d were instantly formed as the sole product by simply adding two equivalents of Pd or Pt square planar complexes to CH2Cl2, CHCl3, or acetone solutions of the cavitand ligands. Positively charged N-methylpyridinium derivatives turned out to be encapsulated via cation-p interactions, the guest being entrapped in the forming capsule. In a subsequent report the authors have demonstrated how metal-induced selfassembly can be tuned by subtle changes in the solvent system [30]. A dynamic equilibrium between the interclipped capsule 23c,d and the intraclipped bowl 24a,b has been observed in nitromethane using [M(dppp)(CF3SO3)2] (M ¼ Pd, Pt) as metal precursors. This equilibrium is totally shifted toward the interclipped capsule in chloroform/methanol, while the intraclipped bowl forms exclusively in the aqueous phase using water-soluble complexes [Pd(en)(NO3)2] [31]. NMR studies clarify the thermodynamic parameters that control the dynamic equilibrium in nitromethane:

j243

244

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Scheme 7.5 Self-assembly of 2þ2 coordination cages 20 and 21 using dinuclear metal precursors.

7.2 Coordination Cages

Scheme 7.6 Solvent-mediated self-assembly of coordination capsules 23a—d and bowls 24a,b.

the conversion from the capsule to the bowl is enthalpically disfavored, but entropically favored. The structure of the interclipped capsule 23d was corroborated by means of X-ray crystallography. Haino et al. synthesized the new octadentate cavitand 25 (Figure 7.6) bearing phenyl-bipyridine groups at the apical position [32]. Cavitand 25 forms a stable dimeric cage 26 via coordination to four Agþ metal centers. Each silver cation coordinates with two bipyridines in a tetrahedral fashion. The size and shape of the large three-dimensional cavity allows encapsulation of rigid guests such as biphenyls 27a,b because of the presence of acidic protons on the guest, creating a CH/p interaction with the resorcinarene cavity. Moreover, when a 1 : 1 mixture of

j245

246

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.6 (a) octadentate bipyridine cavitand 25; (b) the corresponding Ag+ cage 26; (c) encapsulated guests 27a—b and 28; (d) optimized structure of the co-encapsulation complex of 26 with 28 and acetic acid. Reproduced with permission from Ref. [32].

acetic acid and 28 was added to the solution of the capsule, the uncommon co-encapsulation of the guests with surprising selectivity was observed, even in the presence of an equal amount of propionic acid. Out of several possible H-bonded pairs, only the one matching the overall length of the cavity was selected (see Figure 7.6 bottom right). A subsequent detailed thermodynamic study on guest encapsulation within the above-mentioned capsule revealed that not only CH/p interactions between the methyl groups on the guest ends and the aromatic cavity, but also desolvation of the inner cavity, play a key role in guest encapsulation [33].

7.2 Coordination Cages

Scheme 7.7 Self-assembly of the octaanionic cages 30a,b.

Harrison and coworkers introduced four iminodiacetate moieties as chelating ligands at the apical position of resorcinarene-based cavitands (29 in Scheme 7.7). Treatment with CoCl2 under basic conditions led to the formation of cage 30a, with the four octahedrally coordinated Co(II) ions connecting two cavitand scaffolds [34]. Cage 30a, which has been structurally characterized, presents two very interesting features: (a) it is water soluble and (b) the self-assembly process is pH-dependent: at pH values below 2 the carboxyl groups and amine nitrogen atoms are protonated and do not coordinate cobalt(II) ions, but upon addition of a base at pH 5, cage 30a is formed. Taking advantage of the hydrophobic effect, these water-soluble cages are able to include several small liphophilic guests within the cage, such as aromatic molecules, alkanes, alkenes, and haloalkanes [35]. Because of the paramagnetic nature of 30a, large upﬁeld shifts in the proton resonances of the encapsulated guest are observed (30–40 ppm): the cage behaves as an NMR shift reagent for the included guest. Calorimetric studies demonstrated that the most dominant factors in encapsulation selectivity are guest ﬁt and polarity. In most cases, molecules with intermediate polarity are preferred to those that are either non-polar or very polar [36]. Using the same cavitand ligand, the corresponding Fe4 cage 30b has been prepared and structurally characterized [37]. The iron atoms in each complex exhibit a trigonal prismatic geometry. Also, cage 30b is capable of encapsulating organic molecules, moving the chemical shifts of the guests upﬁeld by 20–30 ppm. More recently, the same group described the preparation of a new cage using cavitands substituted

j247

248

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles with bpa ligands and FeCl3 in a 2 : 4 ratio [38]. For the ﬁrst time the cage was characterized by sonic spray mass spectrometry (SSI MS), which allowed the observation of a monocharged ion peak. Thanks to the soft ionization coupled with its simple design, SSI MS provides a powerful tool to characterize such supramolecular cages. 7.2.2.2 Introduction of the Ligands as Bridging Units Concerning the decoration of cavitands at the upper rim, an alternative strategy was described by Gibb and coworkers some years ago [39]. Using this approach in our group, a new series of cavitand ligands was synthesized bearing picolyl groups as bridges between the phenolic OH groups of the resorcinarene scaffold [40]. The synthesis of cavitands 31a,b, was performed by bridging the corresponding resorcinarenes with 4-(a,a-dibromomethyl)pyridine. In this case, unlike the previously described cavitands bearing ligands at the apical positions, each picolyl bridge may assume two different orientations with respect to the cavity, pointing either inward (i) or outward (o) and thus leading to the formation of various isomers. Among these isomers, only the cavitand with four pyridines preorganized in a diverging spatial orientation, i.e. pointing outward from the cavity (oooo isomer), is suitable for cage self-assembly. The correct stereochemical assignment of the oooo isomer has been conﬁrmed by the X-ray crystal structure of cavitand 31b (Figure 7.7a). Moreover, it is possible to evaluate the a angle of the pyridyl ligands relative to the C4v symmetry axis: the estimated value is 42 , close to the ideal value of 45 required for the formation of strainless square-planar complexes. The typical procedure for cage formation is shown in Scheme 7.8: cavitands 31a,b and various metal precursors MLX2 (M ¼ Pd, Pt; L ¼ dppp, en; X ¼ CF3SO3, NO3) were mixed in a 1 : 2 molar ratio at room temperature in solvents like CH2Cl2, CHCl3, acetone, or DMSO, giving cages 32a-h in quantitative yields. In all cases the 1 H NMR spectra showed the formation of a new set of signals, indicative of the presence of a single highly symmetric compound (D4h symmetry). No permanent inclusion of triﬂuoromethanesulfonate counterions was detected by 19 F NMR. Additional proof of cage formation was obtained through ESI and MALDI-TOF mass spectrometry. The X-ray crystal structure of cage 32f (Figure 7.7b) conﬁrms the formation of four square-planar complexes [41]. The four large lateral portals (dimensions 12.3 Å 10.8 Å), allow the entrance/exit of solvent molecules as well as anions, eliminating any possibility of permanent counterion inclusion in solution. In the solid state, one triﬂuoromethanesulfonate anion is included in the cavity, together with two disordered acetone molecules (not shown in Figure 7.7b). The internal volume of the cavity has been estimated as 840 Å3. Several encapsulation experiments with suitably sized guests such as adamantane, 1,2,4,5-tetramethylbenzene, anthracene, pyrene, sodium picrate, and fullerenes in various solvents were carried out. The complexation experiments, monitored by 1 H NMR, were performed at room temperature by forming capsule 32f in the presence of excess guest. The only successful encapsulation experiment performed was the one using methano[60]fullerene derivatives bearing dimethyl and diethyl malonate addends. Evidence for inclusion complexation, provided by 1 H NMR, was conﬁrmed by ESI-MS investigations, which showed the formation of 1 : 1

7.2 Coordination Cages

Figure 7.7 Molecular structure of (a) cavitand 31b and (b) cage 32f, top view as determined by X-ray crystallography. Reproduced with permission from Ref. [41b].

Scheme 7.8 Self-assembly of coordination cages 32a—h.

j249

250

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles fullerene-cage complexes. In collaboration with the group of Diederich and coworkers [41b], the association constant was experimentally determined to be approximately 150 M1 at 298 K in CD2Cl2. The preference for fullerene inclusion is not related to speciﬁc host-guest interactions, but it is driven by dispersion forces and pstacking contacts between the guest and the inner surface of the capsule. An uncommon experiment of cage self-assembly in a liquid-liquid two-phase system has been described. Cavitand 31b (1 equiv) dissolved in 1,1,2,2-tetrachloroethane was exposed to an aqueous solution of Pd(en)(NO3)2 (2.5 equiv) in a test tube. Immediately, a solid formed at the interface, which became more abundant over time. 1 H NMR analysis of the recovered solid indicated the formation of cage 32h in pure form. A MALDI-TOF experiment performed directly on the solid sample conﬁrmed the attribution. This last result is particularly interesting since liquidliquid interfaces represent another potentially useful ﬁeld of action for self-assembly protocols, which has barely been explored so far [42]. The kinetic stability of the Pt cages [43] has been assessed by ESI-MS. A ligandexchange experiment was performed starting from two Pt-cages 32a and 32e, differing by the chains at the lower rim (R ¼ C11H23 and CH2CH2Ph): both cages were separately analyzed by ESI-MS, showing [M-2CF3SO3]2þ, [M-3CF3SO3]3þ, and [M-4CF3SO3]4þ ions. They were then mixed in a 1 : 1 molar ratio at room temperature, giving a spectrum resulting from the simple addition of the signals due to the two homocages. Upon heating the mixture, the ESI-MS spectrum exhibited a third set of peaks belonging to the heterocage in a statistical 1 : 2 : 1 ratio with the signals of the homocages. Another interesting ﬁnding is that the introduction of four methyl groups at the apical positions of cavitand 31 completely switches off the CSA process by hindered rotation of the pyridyl ligands, as in the case of cavitand 14. Our chosen strategy for enlarging the dimensions of the coordination cage lies in deepening the cavitand precursor cavity via introduction of a phenyl spacer between the methylene bridge and the pyridine ligand [44]. For the synthesis of the tetradentate cavitand ligands 33a,b, 4-(a-a-dibromo)tolylpyridine was used as bridging reagent. In this case, only the desired oooo isomer formed for steric reasons. Following the above-mentioned procedure of cage formation, cages 34a–f were obtained in quantitative yields by simply mixing 33a or 33b with MLX2 complexes in a 1 : 2 ratio (Scheme 7.9). The crystal structure of cage 34d revealed the exact position of the eight triﬂuoromethanesulfonate counterions, which are all placed outside the cavity (Figure 7.8), near the metal centers, thanks to the increased distance between the adjacent Pt centers (close Pt-Pt distance ¼ 14.7 Å). It is worth noting that anion inclusion in this class of cages is determined by the distance between the metal centers. When this distance is large enough to permit all counterions to sit close to the metal centers, no inclusion is observed in the solid state (compare Figure 7.7b with Figure 7.8). The new cages reach nanoscopic dimensions, and the estimated internal volume of the cavity is 1800 Å3, more than twice that of the picolyl-bridged analogue 32. A competition experiment has been performed to test the self-recognition properties of cavitands 31b and 33b in the CSA process. After the addition of a

7.2 Coordination Cages

Scheme 7.9 Self-assembly of nanosize coordination cages 34a—f.

stoichiometric amount of [Pd(en)(CF3SO3)2] to an equimolar mixture of 31b and 33b, only the signals belonging to the homocages 32g and 34f were detected. The mismatch between the biting angles of the two cavitands leads to complete selfrecognition during cage self-assembly. In a recent report we described the preparation of cavitand 35, which bears only two tolylpyridyl moieties as bridging ligands in an AC distal position, and its selfassembly behavior with [Re(CO)5Br], which undergoes cis ligand exchange exclusively (Scheme 7.10) [45]. In this case, the self-assembled fac-[Re(CO)3352Br] cage 36 is neutral and kinetically stable at room temperature.

j251

252

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.8 Top view of the molecular structure of 34d. Reproduced with permission from Ref. [44].

In the examples reported so far, the ligand bridging units are oriented outward with respect to the cavity. Dutasta and coworkers reported a case where inwardfacing P ¼ S ligands bridging a resorcinarene skeleton form a coordination cage in the presence of Ag+ cations [46]. A drawback is that the inner space of the cage is highly reduced by the presence of the sulfur atoms. 7.2.3 Trimeric, Tetrameric, and Hexameric Coordination Cages

One of the long-term goals of metal-directed self-assembly is the generation of coordination cages of nanoscale dimensions, capable of conﬁning large objects within their inner space. To widen the inner space of coordination cages, three different strategies have been implemented so far: 1. introduction of rigid spacers between the ligand and the macrocycle (see 34 as an example) 2. use of dinuclear metal precursors (see 20 and 21 as examples) 3. assembly of more than two macrocyclic ligands. This last approach is covered in the present chapter. Beer and coworkers were the ﬁrst to report the metal-directed synthesis of trimeric and tetrameric coordination cages 37 and 38, assembled from dithiocarba-

7.2 Coordination Cages

Scheme 7.10 Self-assembly of fac-[Re(CO)3Br] AC dimeric cage. The relative orientation of the two Br on the Re metal centers has not been determined.

mate (dtc) functionalized cavitands and various metal ions (Figure 7.9) [47]. The coordination preference of the transition metal used dictates the assembly that is adopted. Using group 12 metals such as zinc(II) and cadmium(II), trimeric hexanuclear cavitand assemblies 37a–f form. As shown by X-ray crystallographic studies, the three resorcinarene cups provide the corners for an equilateral molecular triangle with two zinc or cadmium ions positioned at the midpoints of each side. In this case the metal ions coordinate two bidentate dtc moieties and a pyridine molecule. Treatment of the same resorcinarene ligand with copper(II) acetate provided an intermediate copper(II) tetrameric octanuclear complex, whose X-ray crystal structural analysis reveals a ﬂattened tetrahedron shape. Chemical oxidation produced the corresponding copper(III) complexes 38a–c, where an X-ray structure determination displayed a similar tetrahedral arrangement of the cavitand connected to each other by eight copper(III) ions forming approximately square planar complexes. To obtain the analogous gold(III) octanuclear complex 38d, a transmetalation reaction was undertaken on the corresponding copper (II) complex. Strong 1 : 1 complexes are formed between fullerenes (C60 and C70) and hexanuclear assemblies 37a,b,d,e

j253

254

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.9 Trimeric and tetrameric polymetallic architectures 37a—f and 38a—d.

7.2 Coordination Cages

with log Kass ranging from 3 to 6, determined via UV-visible spectroscopy. In particular, the cadmium hexanuclear architectures 37a–c were found to be the superior host systems. ESI-MS provided evidence for the encapsulation of [60]fullerene by the copper(II) octanuclear assemblies. Structural data in conjunction with molecular modeling studies revealed that it is the lability of the copper(II) dithiocarbamate bond which allows guest encapsulation, since the entrance to the nanosized cavity is restricted by its relatively small portals. In both cases it is the electronic character of the dithiocarbamate sulfur donor atoms which, affected by the nature of the metal ion, determines the strength of the binding to the electron-deﬁcient fullerene guest species. Another example of trimeric polymetallic architectures has been described by Stang and coworkers starting from bispyridyl-cavitand ligands and linear bimetallic precursors such as 1,4-[bis((PEt3)2Pt(NO3))2-benzene] and [4,40 -bis((PEt3)2Pt(NO3))2-biphenyl] in a CD3NO2/D2O two-phase solvent system (Figure 7.10) [28]. The four nanoscopic metallacyclic cavitand complexes 39a–d were characterized by a combination of NMR analysis and ESI mass spectrometry. In particular, the 31 P and 1 H NMR spectra of these assemblies suggest highly symmetrical architectures on the NMR timescale. Atwood and coworkers developed a different strategy. They used calix[4]arenes and pyrogallol[4]arenes [48] directly as multidentate ligands, avoiding further functionalization of the macrocycles. By simply deprotonating the sulfonic and phenolic OHs, they obtained the corresponding chelating ligands, capable of complexing metal ions. They described the formation of nanometer-scale spheres and tubules assembling p-sulfonatocalix[4]arene building blocks by the addition of pyridine-N-oxide and lanthanide ions [49]. Crystallographic studies revealed how metal ion coordination and substrate recognition direct the formation of these assemblies. In particular, from a 2 : 2 : 1 aqueous solution of p-sulfonatocalix[4]arene sodium salt, pyridine-N-oxide and La(NO3)36H2O, spherical supermolecules form that contain 12 calixarene anions placed at the vertices of an icosahedron (Figure 7.11a). On the other hand, mixing the three components in a 2 : 8 : 1 molar ratio, tubular assemblies form which show a chiral helical assembly along the length of the tube. In another article, the same group reported the formation of a large neutral hexameric pseudo-spherical coordination capsule assembled from 6 pyrogallol[4]arene ligands and 24 Cu(II) metal ions [50]. A single-crystal X-ray analysis shows that the centroids of the six pyrogallol[4]arene ligands reside on the vertices of a hypothetical octahedron and enclose a cavity of 1200 Å3. Amazingly, this coordination capsule is structurally analogous to its hydrogen-bonded counterpart (Figure 7.11b, c). These hydrogen-bonded assemblies function as a supramolecular blueprint during metal ion addition, preserving the desired structural framework in response to the interchange between the hydrogen-bonded networks for metal coordination. After the insertion of 24 Cu(II) metal ions, a combination of 96 Cu-O coordination bonds and 24 intramolecular hydrogen bonds binds the framework together. A different neutral hexameric coordination cage structure was assembled from six either C-propyl- or C-pentyl-pyrogallol[4]arene units and 12 Ga(III) ions [51]. In both cases the centroids of the macrocyclic ligand components reside on the vertices of a squeezed octahedron enclosing a space of approximately 1150 Å3. Only a total of 20

j255

256

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.10 Trimeric assemblies of bispyridyl cavitand ligands 39a—d.

7.2 Coordination Cages

Figure 7.11 (a) Spherical assembly of sulfonatocalix[4]arene sodium salt; (b) hydrogen- and (c) metal-directed selfassembly of six pyrogallol[4]arenes. Reproduced with permission from Refs. [49,50].

intramolecular hydrogen bonds remain after insertion of gallium atoms which bind the framework together. 7.2.4 Self-assembly of Coordination Cages on Surfaces

The clean, precise, and reversible modiﬁcation of surfaces represents an important opportunity to expand the application range of self-assembly. In particular, the thermodynamic control and reversibility of CSA can be exploited for the error-free generation of three-dimensional architectures directly on surfaces. Our group, in collaboration with Reinhoudt and coworkers [52], has recently reported an intriguing example of the generation of coordination cages directly on surfaces by using self-assembled monolayers (SAMs) as molecular platforms. First,

j257

258

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Scheme 7.11 Self-assembly leading to the formation of cages on gold.

monolayers of 40 were prepared by adsorption of the thioether chains on gold surfaces. The layers obtained were used to perform the metal-induced capping reactions directly on the SAMs through the addition of a solution of 8a and [Pd(dppp)(CF3SO3)2] (cage 41a) in the ﬁrst case or of 8a and [Pt(dppp)(CF3SO3)2] (cage 41b) in the second case (Scheme 7.11). The formation of cages 41a,b on SAMs was proved by contact-angle measurements, electrochemistry, X-ray photoelectron spectroscopy (XPS), and atomic force microscopy (AFM). A series of control experiments were performed in order to prove that the self-assembly of the cages occurs on the surface of the SAMs via metal coordination. Treatment of the SAM of 41a with triethylamine shifted the equilibrium toward the formation of [Pd(dppp) (NEt3)2(CF3SO3)2] plus uncomplexed cavitands. The complete disassembly of the cages on the SAM was shown by electrochemistry measurements. Formation of the

Figure 7.12 CSA on microcontact-printed surfaces. Reproduced with permission from Ref. [52].

7.2 Coordination Cages

j259

260

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.13 Thioether-functionalized cavitand ligand and the corresponding homocage 43 and heterocage 44.

molecular cages was monitored directly by AFM. For this purpose, microcontactprinted templates were created using standard soft lithography methods [53]. These templates showed the presence of very regular features over large areas (Figure 7.12). Stripes with a lower height correspond to areas exposing bare gold, whilst stripes with a higher topographic proﬁle are covered by the SAM of 11-mercaptoundecanol (Figure 7.12a). After deposition of a monolayer of cavitand 40 on the bare gold grooves, AFM was used to directly measure the differences in the height proﬁle (Figure 7.12b). The height difference of 2.6 nm is consistent with the formation of a monolayer of 40. Self-assembly of the molecular cages on the printed layers was performed using the same procedure as for the full monolayers, leading to a total height difference of 4.2 nm from the bare gold grooves (Figure 7.12c), a value which is consistent with the calculated dimensions of the cage. A different approach has been undertaken when highly ordered monolayers of cavitands do not form. This typically happens when the cross-sectional area of the

Figure 7.14 Schematic representation and AFM section analysis of a MU SAM with inserted cavitand 42 (a), heterocage 44 (b), and homocage 43 (c). Reproduced with permission from Ref. [54].

7.2 Coordination Cages

j261

262

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.15 3D-AFM image of MU SAM with inserted homocage 43.

cavitand head exceeds that of the underlying feet, as in the case of 42 (Figure 7.13) [54]. The CSA on disordered full SAMs of 42 do not work properly because of the interdigitation of the tolylpyridyl ligands. As an alternative, the insertion of cavitand 42 into a monolayer of 11-mercaptoundecanol (MU) was achieved by soaking SAMs of thiol in a 0.25-mM solution of 42 in dichloromethane. Extensive washing with large amounts of solvents guaranteed the removal of all physisorbed material. AFM experiments and proﬁle analysis showed the presence of individual round entities which protrude from the ﬂat by approximately 2 nm, which is the expected height difference between cavitand 42 and MU (Figure 7.14a). The next step was the assembly of the heterocage 44 by exposing the SAM of 42 in MU to a solution of cage 34e, bearing hexyl tails at the lower rim. The formation of heterocage 44 on gold was thus obtained through ligand exchange. In this case, AFM analysis revealed the presence of two different protruding objects: (a) those with a height of about 4–5 nm, corresponding to the heterocage, and (b) those with a height of about 2 nm due to the presence of free cavitand ligands on gold (Figure 7.14b). The outcome of this analysis suggests that both uncapped cavitands and heterocages are present because the very dilute conditions employed do not allow a complete ligand exchange. For the formation of coordination cages on gold, another strategy has been explored which leads to the insertion of preformed homocage 43 into a monolayer of MU. Homocage 43 has been directly inserted by soaking a clean MU SAM in a dichloromethane solution of the preformed cage for 1 h at room temperature (Figures 7.14c and 7.15). Extensive rinsing is necessary to ensure the removal of physisorbed material. AFM analysis showed the presence of objects protruding 5.1–5.3 nm from the MU SAM, in accordance with the expected protrusion length of the cages. As in the previously described case of cyano-cages grafted on gold, a control experiment was carried out, consisting in the disassembly of cage 43 immobilized on gold. By exposing a gold substrate with inserted homocages to a solution of triethylamine, more than half of the cages were disassembled, as shown in the AFM analysis. Another topic related to the self-assembly of coordination cages in the solid state has been described by Harrison and coworkers [55]. Octaanionic cages 30a with four Ca2þ countercations form grids in the solid state and pack into porous materials resembling zeolites. The carboxylate groups on the four sides of cage 30a can bind to more than one metal ion. In particular, the calcium ions coordinate with carboxylates from two

7.2 Coordination Cages

Figure 7.16 Structure of the Ca-assembled material: (a) top view and (b) side view. Reproduced with permission from Ref. [55b].

cages, creating a bridge between them. Thus, octaanionic 30a arranges into sheets of cages linked together by Ca2þ. Voids are formed between neighboring cages, and water molecules bound to calcium ions surround them. Regarding the placement of the next layers, the cages of one layer are positioned above the cation bridges of another layer. The third layer lies directly over the ﬁrst layer, creating an ABAB cage packing arrangement, and this results in the formation of rhombic pores of width 13 Å and length 21 Å. The channels connecting the pores lie between the layers (Figure 7.16). These pores and channels contain water, which can be removed by thermal desorption. In fact, when the temperature of the material is raised from 25 to 250 C, a total weight loss of 20% is observed, corresponding to a 40 : 1 water-to-cage molar ratio in the pristine material. After releasing water, this material is able to regain the amount of water that was lost by cooling in the presence of a stream of water vapor. Moreover, pore and channel size and shape can be tuned by using different alkaline-earth cations (Mg2þ, Ca2þ, Sr2þ, Ba2þ) having different sizes and coordination numbers. The heat of water adsorption ranges from 46 to 52 kJ mol1, values similar to those of silica-containing zeolites. As recently reported, the properties of these materials can be further differentiated through the incorporation of transition (i.e. Co2þ) or alkaline cations (i.e. Kþ) into the channels of barium-linked materials composed of metal-assembled cages. 7.2.5 Self-assembly of Multitopic Macrocyclic Complexes

While the design and formation of self-assembled coordination cages described in the previous sections have received much attention, the corresponding metal-directed self-assembly of open multitopic receptors has been almost neglected so far. One reason is related to the need for partially derivatized precursors, which are synthetically more demanding than the corresponding fully derivatized ones. The ﬁrst approach to the formation of a ditopic cavitand complex has been reported by Sorrell et al. [56]. They identiﬁed the tri-bridged resorcinarene as a

j263

264

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.17 Cis-ditopic resorcinarene complexes 45a—e and the molecular structure of complex 45a. Reproduced with permission from Ref. [56].

Figure 7.18 Tolylpyridyl monodentate cavitand ligand 46.

7.2 Coordination Cages

Figure 7.19 Cis-ditopic cavitand complex 47 and its crystal structure (side view). Reproduced with permission from Ref. [57].

suitable monodentate ligand in its monodeprotonated form. Two of these macrocyclic monodentate ligands have been linked using M(II) metal precursors having bipyridines or bisimidazoles as chelating ligands (Figure 7.17). In the case of copper complex 45a, the crystal structure revealed that the two macrocycles are not facing each other, but rather lie side by side, ruling out any possibility of cooperative guest binding. More recently our group has tackled this problem in a different way [57]. The starting point was again a tri-bridged resorcinarene, in which the last pair of free OHs has been used to introduce a single pyridine (not shown) or tolylpyridyl ligand (46a,b in Figure 7.18) as bridging unit. These monodentate cavitands are versatile ligands for the formation of ditopic and tetratopic complexes either charged or neutral. The formation of charged ditopic complexes with the two cavities facing each other is straightforward, employing the same metal precursors and conditions used in the CSA protocols (see for example complex 47 in Figure 7.19 where [Pt (dppp)(CF3SO3)2] has been used).

j265

Figure 7.20 Tetratopic cavitand complex 48.

Figure 7.21 Trans-ditopic cavitand complex 49 and its molecular structure as determined by X-ray crystallography (side view). Reproduced with permission from Ref. [57].

7.2 Coordination Cages

Figure 7.22 Cis-ditopic neutral cavitand complex 50.

Using [Pd2(tppb)(CF3SO3)4] as a dinuclear metal precursor, the corresponding tetratopic complex 48 was assembled in quantitative yields (Figure 7.20). The only drawback of such complexes is related to their limited kinetic stability in solution, which can hamper their use as multitopic receptors, particularly for charged guests that increase the ionic strength of the medium. We therefore turned our attention to the self-assembly of neutral complexes, which are kinetically stable in solution, starting from the same cavitand ligands [57]. The use of cis-[Pt(CH3CN)2Cl2] and cis-[Pt(DMSO)2Cl2] as metal precursors led to the formation of the trans complex 49 in both cases (Figure 7.21). The unexpected stereochemistry of the ﬁnal product is the result of the cis-to-trans isomerization of the intermediate monofunctionalized complexes. To overcome this problem we switched to [Re(CO)5Br] as the metal precursor, because it is one of the few octahedral complexes that undergoes an exclusive cis ligand exchange, because of the remarkable trans effect of the carbonyl ligands. The replacement of two equatorial cis-CO groups gave the desired neutral

j267

268

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Scheme 7.12 Self-assembly of tetratopic Pd/Pt complex 52 and the other two possible cyclic structures. In the top views sketched in the scheme each sphere corresponds to two cavitands, one over the other.

7.2 Coordination Cages

Scheme 7.13 Self-assembly of ditopic container 54 and inclusion of guests (alkylammonium salts).

fac-[(CO)346b2Br] ditopic complex 50 (Figure 7.22). This last protocol was found to be the best one for the generation of kinetically inert, ditopic receptors. In Section 7.2.2 the self-assembly of bidentate AC cavitand 35 using [Re(CO)5Br] as the metal precursor to obtain the dimeric capsule 36 was described. The corresponding AB bidentate cavitand 51 has been reacted with dinuclear metal precursors [M2(tppb)(CF3SO3)4] (M ¼ Pt, Pd) by mixing the two building blocks in a 2 : 1 ratio [45]. At this level of complexity, the self-assembly can lead to three possible cyclic structures (tetratopic, hexatopic, and octatopic cavitand complexes, Scheme 7.12). Out of them, only the entropically favored tetratopic complex 52 forms, as predicted by molecular modeling and demonstrated by PGSE NMR and ESI-MS experiments. The potential of ditopic cavitand complexes to bind ditopic guests has been demonstrated by Rebek and coworkers using cavitand 53 as monodentate ligand (Scheme 7.13) [58]. Cavitand 53 combines the advantages of a deep vase-like cavity obtained through a network of hydrogen bonds and the presence of a tolylpyridyl group able to coordinate to the metal center. The ditopic complex 54 forms quanti-

j269

270

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles

Figure 7.23 Calix[5]arene ditopic receptor 55.

tatively by combining 53 and [Pt(dppp)(CF3SO3)2] in a 2 : 1 ratio; it has nanoscale dimensions and is able to reversibly bind alkylammonium guests. The presence of two receptors facing each other has a twofold effect: 1. Reduction of the exchange rate of the alkyldiammonium guest B compared to the alkylammonium A 2. The inclusion of B takes place with a positive cooperative effect gained from the preorganization of the host 54. An interesting example of a self-assembled ditopic receptor (55) has been developed by Fukazawa and coworkers starting from two calix[5]arenes with bipyridine units at the upper rim (Figure 7.23) [59]. In a previous article, the same authors

Figure 7.24 2,20 -bipyridine-bridged biscavitand 56 and its Agþ complex 57.

7.3 Conclusion

reported that the covalent linking of two calix[5]arene units brings strong binding to C60 and C70 within the cavity formed by the two calixarene scaffolds [60]. Also, in this case, metal-directed self-assembly through the addition of 0.5 equivalents of Ag+ leads to the formation of the ditopic receptor 55 capable of forming a 1 : 1 inclusion complex with fullerene molecules (C60 and C70). Examples of ditopic and tetratopic macrocycles have been described by Stang and coworkers by combining the calixarene chemistry with the metal-directed self-assembly of molecular squares [61]. In this case the calixarene units coordinate to the Pt through their P(Ph)2 ligands at the lower rim and assume a diverging spatial orientation in the macrocycle. These calixarene metallamacrocycles are used in transport experiments to carry sulfonate salts from one aqueous phase into another through chloroform. Allosteric behavior in a ditopic cavitand receptor has been induced via metal coordination [62]. Addition of Agþ ions to a solution of a 2,20 -bipyridine-bridged biscavitand 56 switches the molecule from an S-shaped to a C-shaped conformation with the two cavities facing each other. At the same time two bis-cavitand molecules are brought together to give an overall tetratopic complex 57 (Figure 7.24).

7.3 Conclusion

The stimulus behind the rapid development of molecular containers is their potential application as molecular devices. The metal-directed assembly of synthetic macrocycles represents a valuable tool for the generation of such container molecules avoiding the problems related to the synthesis of their purely organic counterparts, which often is time consuming and leads to the formation of kinetically distributed products. The results described in this chapter demonstrate the vast and almost unlimited possibilities of creating new molecular architectures constructed in simple self-assembly processes starting from calixarene and cavitand scaffolds. The ﬁnal shape of the metal-directed nanoscopic structures is deﬁned by two main factors: (a) the rational design of the macrocyclic building blocks through the correct orientation of their ligand groups, and (b) the accurate choice of the metal center and its coordination geometry. In this way the metal-directed self-assembly of the two components produces architectures having predeﬁned and predictable shapes and sizes in quantitative yields under mild conditions. Among the several challenges and opportunities looming ahead, in our opinion two are worth pursuing. One is the self-assembly of coordination cages endowed with molecular recognition properties determined by speciﬁc interactions and not by hydrophobic or shape-driven inclusion. This requires the decoration of the inner space of the coordination cages with suitable functional groups, which must be spatially organized and inert toward metal coordination. The other challenge concerns the extension of the self-assembly protocol to technologically interesting surfaces, like silicon, in order to build hybrid inorganic-organic structures featuring selective inclusion and other useful properties. The generation of such complex

j271

272

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles organic architectures on silicon will be pivotal in developing new integrated devices presenting peculiar optical, magnetic, and sensing properties.

7.4 Experimental: Selected Procedures 7.4.1 Tetrapicolyl-bridged Cavitand 31a

(This synthetic approach has been applied for the synthesis of cavitand 31b from the corresponding CH2CH2Ph-footed resorcinarene.) A mixture of undecyl-footed resorcinarene (0.707 g, 0.64 mmol), K2CO3 (1.06 g, 7.67 mmol), and 4-(dibromomethyl)pyridine [40] (1.61 g, 6.40 mmol) in 30 mL of dry N,N-dimethylacetamide (DMA) was stirred in a sealed tube at 80 C for 15 h. The reaction mixture was then cooled, poured into a saturated Na2CO3 solution, and extracted with CH2Cl2. Removal of the CH2Cl2 gave a black solid that was puriﬁed in two steps: ﬁrst by column chromatography (SiO2, CH2Cl2/EtOH from 9 : 1 to 1 : 1), then on silica gel plates (CH2Cl2/EtOH 9 : 1) to give 90 mg (0.06 mmol) of cavitand 31a (yield 10%). 7.4.2 Tetracyano Cavitand 40

(This synthetic approach has been applied for the synthesis of cavitands 8a–c from the corresponding tetrabromo derivatives.) Cavitand 40 was prepared from the corresponding tetrabromo derivative [63]. A mixture of tetrabromo cavitand (1.8 g, 0.865 mmol) and of CuCN (0.767 g, 8.56 mmol) in NMP was reﬂuxed for 24 h at 200 C. After cooling, a mixture of FeCl36H2O (2.32 g, 8.58 mmol) and 1 N HCl (8.5 mL, 8.5 mmol) was added, and the rection mixture was heated to 75 C for 30 min. After cooling, the solvent was evaporated, and the residue was dissolved in dichloromethane and washed with NH4Cl, brine, and H2O, three times each respectively. The organic phase was dried with MgSO4 and subsequently concentrated under reduced pressure. Puriﬁcation was performed via column chromatography (SiO2: hexane/ethyl acetate 7 : 1). After recrystallization from methanol, 40 (1.1 g, 0.588 mmol) was obtained as a white solid (yield 68%). 7.4.3 AC-dibridged Tolylpyridyl Cavitand 35

(This synthetic approach has been applied for the synthesis of AB-dibridged tolylpyridyl cavitand 51 from the corresponding AB-dimethylene-bridged resorcinarene.) 4-(a,a-Dibromotolyl)pyridine (0.3 g, 0.92 mmol) and K2CO3 (0.35 g, 2.5 mmol) were added under nitrogen to a solution of the corresponding AC-dimethylenebridged resorcinarene (0.19 g, 0.31 mmol) [45] in dry DMA (N,N-dimethylaceta-

Abbreviations

mide, 15 mL). The mixture was stirred in a sealed tube at 80 C for 16 h. The reaction was quenched by addition of 10 mL water, and the resulting mixture was extracted with 15 mL CH2Cl2. The organic layer was washed with water (3 · 15 mL), dried over Na2SO4, and evaporated. The material obtained was puriﬁed by column chromatography (SiO2: hexane/acetone/EtOH 7 : 2 : 1) to give cavitand 35 (0.12 g, 0.13 mmol, out-out isomer) as a white solid (yield 42%). 7.4.4 fac-Br(CO)3Re AC Ditopic Cavitand Complex 36

Cavitand 35 (41 mg, 4.29 · 102 mmol) was added to a solution of [Re(CO)5Br] (14.0 mg, 3.43 · 102 mmol) in CHCl3 and heated at reﬂux for 24 h. After solvent evaporation, the product was puriﬁed by column chromatography (SiO2: CHCl3/ ethyl acetate 6 : 4) to give the fac-ditopic cavitand complex 36 (30 mg, 6.9 · 103 mmol, yield 40%). 7.4.5 Tetratopic Cavitand Complex 48

Tetratopic cavitand complex 48 was assembled by mixing cavitand 46b with the dinuclear complex [Pd2(tppb)(CF3SO3)4] [57] in a 4 : 1 molar ratio at room temperature in 10 mL of a CHCl3/CH3NO2/CH3CN 7 : 2 : 1 mixture. After stirring for 15 min, the desired complex 48 was obtained in quantitative yield after removal of the solvent under vacuum.

Acknowledgments

Our own work described in this chapter has been supported over the years by the following grants: PRIN 2001 and PRIN 2003 (MURST-Rome), FIRB-Nanoorganization of Materials with Magnetic and Optical Properties, European Network of Excellence MAGMANET.

Abbreviations

dppp dppb DABCO DaniF CSA en bpa SSI MS dtc

1,3-bis(diphenylphosphino)propane 1,4-bis(diphenylphosphino)butane 1,4-diazabicyclo[2.2.2]octane N,N0 -di-p-anisylformamidinate cage self-assembly ethylenediamine bis(2-pyridylmethyl)amine sonic spray mass spectrometry dithiocarbamate

j273

274

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles SAM XPS AFM MU tppb oooo DMSO acac py NMP DMA MALDI-TOF ESI

self-assembled monolayer X-ray photoelectron spectroscopy atomic force microscopy 11-mercaptoundecanol 1,2,4,5-tetrakis(diphenylphosphino)benzene out,out,out,out (the four ligands point outward) dimethyl sulfoxide acetylacetonate pyridine 1-methyl-2-pyrrolidinone N,N-dimethylacetamide matrix assisted laser desorption ionization-time-of-ﬂight mass spectrometry electrospray ionization mass spectrometry

References 1 Gutsche, C.D. (1989) Calixarenes, Royal Society of Chemistry, Cambridge. 2 Cram, D.J. and Cram, J.M. (1994) Container Molecules and Their Guests, Royal Society of Chemistry, London. 3 Rebek, J., Jr. (1990) Angew. Chem., 102, 261–272; Angew. Chem. Int. Ed., 29, 245– 255. 4 Collet, A., (Eds. J.L. Atwood , J.E.D. Davis , D.D. MacNicol and F. Vögtle) (1996) Comprehensive Supramolecular Chemistry, Vol. 2, Pergamon, Oxford, pp.325–365. 5 (a) Conn, M.M. and Rebek, J., Jr. (1997) Chem. Rev., 97, 1647–1668. (b) for the use of ionic interactions in the selfassembly of molecular containers see: Corbellini, F., Fiammengo, R., Timmerman, P., Crego-Calama, M., Versluis, K., Heck, A.J.R., Luyten, I. and Reinhoudt, D.N. (2002) J. Am. Chem. Soc., 124, 6569–6575. 6 Leininger, S., Olenyuk, B. and Stang, P. J. (2000) Chem. Rev., 100, 853–908. 7 Gibb, C.L.D. and Gibb, B.C. (2004) J. Am. Chem. Soc., 126, 11408–11409. 8 For two selected examples see: (a) Chen, J., Körner, S., Craig, S.L., Rudkevich, D. M. and Rebek, J., Jr. (2002) Nature, 415, 385–386. (b) Koblenz, T.S., Dekker, H.L.,

9

10

11 12

13 14 15 16

de Koster, C.G., van Leeuwen, P.W.N.M. and Reek, J.N.H. (2006) Chem. Commun., 1700–1702. Rudkevich, D.M. (2004) Angew. Chem., 116, 568–581. Angew. Chem. Int. Ed., 43, 558–571. Hof, F., Craig, S.L., Nuckolls, C. and Rebek, J., Jr. (2002) Angew. Chem., 114, 1556–1578. Angew. Chem. Int. Ed., 41, 1489–1508. Warmuth, R. (2001) Eur. J. Org. Chem., 423–437. For the few examples of controlled guest delivery from hemicarceplexes see: (a) Piatnitski, E.L. and Deshayes, K.D. (1998) Angew. Chem., 110, 1022–1024. (1998) Angew. Chem. Int. Ed., 37, 970– 972. (b) Ro, S., Rowan, S.J., Pease, A.R., Cram, D.J. and Stoddart, J.F. (2000) Org. Lett., 2, 2411–2414. Rudkevich, D.M. (2002) Bull. Chem. Soc. Jpn., 75, 393–413. Service, R.F. (2002) Science, 295, 2398– 2399. Fujimoto, K. and Shinkai, S. (1994) Tetrahedron Lett., 35, 2915–2918. Ikeda, A., Yoshimura, M., Tani, F., Naruta, Y. and Shinkai, S. (1998) Chem. Lett., 587–588.

References 17 Ikeda, A., Yoshimura, M., Udzu, H., Fukuhara, C. and Shinkai, S. (1999) J. Am. Chem. Soc., 121, 4296–4297. 18 Ikeda, A., Udzu, H., Yoshimura, M. and Shinkai, S. (2000) Tetrahedron, 56, 1825–1832. 19 Zhong, Z., Ikeda, A., Ayabe, M., Shinkai, S., Sakamoto, S. and Yamaguchi, K. (2001) J. Org. Chem., 66, 1002– 1008. 20 Baldini, L., Ballester, P., Casnati, A., Gomila, R.M., Hunter, C.A., Sansone, F. and Ungaro, R. (2003) J. Am. Chem. Soc., 125, 14181–14189. 21 Cotton, F.A., Lei, P., Lin, C., Murillo, C. A., Wang, X., Yu, S.-Y. and Zhang, Z.-X. (2004) J. Am. Chem. Soc., 126, 1518– 1525. 22 Jacopozzi, P. and Dalcanale, E. (1997) Angew. Chem., 109, 665–667. Angew. Chem. Int. Ed., 36, 613–615. 23 Fochi, F., Jacopozzi, P., Wegelius, E., Rissanen, K., Cozzini, P., Marastoni, E., Fisicaro, E., Fokkens, R. and Dalcanale, E. (2001) J. Am. Chem. Soc., 123, 7539– 7552. 24 Zuccaccia, D., Pirondini, L., Pinalli, R., Dalcanale, E. and Macchioni, A. (2005) J. Am. Chem. Soc., 127, 7025–7032. 25 Cuminetti, N., Ebbing, M.H.K., Prados, P., de Mendoza, J. and Dalcanale, E. (2001) Tetrahedron Lett., 42, 527–530. 26 Kobayashi, K., Yamada, Y., Yamanaka, M., Sei, Y. and Yamaguchi, K. (2004) J. Am. Chem. Soc., 126, 13896–13897. 27 Yamanaka, M., Yamada, Y., Sei, Y., Yamaguchi, K. and Kobayashi, K. (2006) J. Am. Chem. Soc., 128, 1531–1539. 28 Jude, H., Sinclair, D.J., Das, N., Sherburn, M.S. and Stang, P.J. (2006) J. Org. Chem., 71, 4155–4163. 29 Park, S.J. and Hong, J.-I. (2001) Chem. Commun., 1554–1555. 30 Park, S.J., Shin, D.M., Sakamoto, S., Yamaguchi, K., Chung, Y.K., Lah, M.S. and Hong, J.-I. (2005) Chem. Eur. J., 11, 235–241. 31 Park, S.J., Shin, D.M., Sakamoto, S., Yamaguchi, K., Chung, Y.K., Lah, M.S.

32

33 34

35

36

37 38

39

40

41

42

43

44

and Hong, J.-I. (2003) Chem. Commun., 998–999. Haino, T., Kobayashi, M., Chikaraishi, M. and Fukazawa, Y. (2005) Chem. Commun., 2321–2323. Haino,T., Kobayashi, M. and Fukazawa, Y. (2006) Chem. Eur. J., 12, 3310–3319. Fox, O.D., Dalley, N.K. and Harrison, R. G. (1998) J. Am. Chem. Soc., 120, 7111– 7112. Fox, O.D., Leung, J.F.-Y., Hunter, J.M., Dalley, N.K. and Harrison, R.G. (2000) Inorg. Chem., 39, 783–790. Harrison, R.G., Burrows, J.L. and Hansen, L.D. (2005) Chem. Eur. J., 11, 5881–5888. Fox, O.D., Dalley, N.K. and Harrison, R. G. (1999) Inorg. Chem., 38, 5860–5863. Gardner, J.S., Harrison, R.G., Lamb, J.D. and Dearden, D.V. (2006) New J. Chem., 30, 1276–1282. Xi, H., Gibb, C.L.D., Stevens, E.D. and Gibb, B.C. (1998) Chem. Commun., 1743–1744. Pirondini, L., Bertolini, F., Cantadori, B., Ugozzoli, F., Massera, C. and Dalcanale, E. (2002) Proc. Natl. Acad. Sci. U. S. A., 99, 4911–4915. (a) Pirondini, L. and Dalcanale, E. (2004) Dekker Encyclopedia of Nanoscience and Nanotechnology 3415–3430. (b) Pirondini, L., Bonifazi, D., Cantadori, B., Braiuca, P., Campagnolo, M., De Zorzi, R., Geremia, S., Diederich, F. and Dalcanale, E. (2006) Tetrahedron, 62, 2008–2015. Bowden, N., Terfort, A., Carbeck, J. and Whitesides, G.M. (1997) Science, 276, 233–235. Dynamic interconversion of similar square-planar Pd(II) complexes with pyridyl cycloveratrylenes has been shown by Shinkai et al. to occur at room temperature: Zhong, Z., Ikeda, A., Shinkai, S., Sakamoto, S. and Yamaguchi, K. (2001) Org. Lett., 3, 1085– 1087. Pinalli, R., Cristini, V., Sottili, V., Geremia, S., Campagnolo, M., Caneschi,

j275

276

j 7 Supramolecular 3D Architectures by Metal-directed Assembly of Synthetic Macrocycles 45

46

47

48

49

50

51

52

A. and Dalcanale, E. (2004) J. Am. Chem. Soc., 126, 6516–6517. Menozzi, E., Busi, M., Massera, C., Ugozzoli, F., Zuccaccia, D., Macchioni, A. and Dalcanale, E. (2006) J. Org. Chem., 71, 2617–2624. Bibal, B., Tinant, B., Declercq, J.-P. and Dutasta, J.-P. (2002) Chem. Commun., 432–433. (a) Fox, O.D., Drew, M.G.B. and Beer, P. D. (2000) Angew. Chem., 112, 139–144. Angew. Chem. Int. Ed., 39, 136–140. (b) Fox, O.D., Drew, M.G.B., Wilkinson, E.J.S. and Beer, P.D. (2000) Chem. Commun., 391–392. (c) Fox, O.D., Cookson, J., Wilkinson, E.J.S., Drew, M. G.B., MacLean, E.J., Teat, S.J. and Beer, P.D. (2006) J. Am. Chem. Soc., 128, 6990–7002. Cometti, G., Dalcanale, E., Du vosel, A. and Levelut, A.-M. (1990) J. Chem. Soc. Chem. Commun., 163–165. (a) Orr, G.W., Barbour, L.J. and Atwood, J.L. (1999) Science, 285, 1049–1052. (b) Dalgarno, S.J., Atwood, J.L. and Raston, C.L. (2006) Chem. Commun., 4567–4574. McKinlay, R.M., Cave, G.W.V. and Atwood, J.L. (2005) Proc. Natl. Acad. Sci. U. S. A., 102 (17), 5944–5948. (a) McKinlay, R.M., Thallapally, P.K., Cave, G.W.V. and and Atwood, J.L., (2005) Angew. Chem., 117, 5879–5882; Angew. Chem. Int. Ed., 44, 5733–5736. (b) McKinlay, R.M., Thallapally, P.K. and Atwood, J.L. (2006) Chem. Commun., 2956–2958. Levi, S.A., Guatteri, P., van Veggel, F.C.J. M., Vancso, G.J., Dalcanale, E. and Reinhoudt, D.N. (2001) Angew. Chem.,

53

54

55

56 57

58 59

60

61

62 63

113, 1945–1948; Angew. Chem. Int. Ed., 40, 1892–1896. Xia, Y. and Whitesides, G.M. (1998) Angew. Chem., 110, 568–594. Angew. Chem. Int. Ed., 37, 550–575, and references within. Menozzi, E., Pinalli, R., Speets, E.A., Ravoo, B.J., Dalcanale, E. and Reinhoudt, D.N. (2004) Chem. Eur. J., 10, 2199–2206. (a) Harrison, R.G., Dalley, N.K. and Nazarenko, A.Y. (2000) Chem. Commun., 1387–1388. (b) Harrison, R.G., Fox, O. D., Meng, M.O., Dalley, N.K. and Barbour, L.J. (2002) Inorg. Chem., 41, 838–843. (c) Harrison, R.G., Burrows, J. L. and Dalley, N.K. (2003) Inorg. Chim. Acta, 351, 399–403. Sorrell, T.N., Pigge, F.C. and White, P.S. (1994) Inorg. Chem., 33, 632–635. Menozzi, E., Busi, M., Ramingo, R., Campagnolo, M., Geremia, S. and Dalcanale, E. (2005) Chem. Eur. J., 11, 3136–3148. Menozzi, E. and Rebek, J., Jr. (2005) Chem. Commun., 5530–5532. Haino, T., Araki, H., Yamanaka, Y. and Fukazawa, Y. (2001) Tetrahedron Lett., 42, 3203–3206. Haino, T., Yanase, M. and Fukazawa, Y. (1998) Angew. Chem., 110, 1044–1046. (1998) Angew. Chem. Int. Ed., 37, 997–998. Stang, P.J., Cao, D.H., Chen, K., Gray, G. M., Muddiman, D.C. and Smith, R.D. (1997) J. Am. Chem. Soc., 119, 5163– 5168. Lützen, A., Haß, O. and Bruhn, T. (2002) Tetrahedron Lett., 43, 1807–1811. van Velzen, E.U.T. (1994) Ph.D thesis, University of Twente (The Netherlands).

j277

8 New Properties and Reactions in Self-assembled M6L4 Coordination Cages Makoto Fujita and Michito Yoshizawa

8.1 Introduction

In 1990, our group ﬁrst reported the quantitative self-assembly of the Pd(II)cornered square molecule 1 in water [1]. The molecular design of the square within which a transition metals 90-degree coordination angle occupies every corner of the square has since been conﬁrmed by a number of groups studying a variety of squares, e.g., organic soluble squares [2], luminescent squares [3], porphyrin squares [4], chiral squares [5], and crown or ferrocene-conjugated squares [6]. These molecules, called molecular squares, consist of transition metal corners linked together by linear organic ligands.

The extension of the basic two-dimensional design of the squares and related polygons into three-dimensional systems led us to explore the self-assembly of Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

278

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

metal-cornered polyhedra. In this way, the two-dimensional structure of Pd(II)-cornered square 1 was extended into an octahedral coordination cage 2 with composition M6L4 [7]. In this compound, the six vertices of an octahedron are occupied by cis endcapped Pd(II), whereas the eight triangular faces of the hypothetical octahedron are alternately occupied by four exo-tridentate triangular ligands. At each Pd(II) corner, the bite angle of the two pyridyl groups is approximately 90 degrees, causing almost no distortion in this cage framework. Hence, this compound is self-assembled quantitatively by mixing the metal and the ligand in water. Thanks to this facile synthetic procedure, compound 2 is now produced on a large scale as a commercial product.

Metal-cornered cages with the same symmetry have subsequently been reported by many groups, and the self-assembly of polygons and polyhedra using di- or multidentate exotopic ligands and metal hinges is now recognized as a very efﬁcient and versatile way for the production of large hollow coordination objects. In this article, a focus will be on the hollow complex 2 [7], which is one of the ﬁrst examples of a self-assembled cage. The self-assembly of the cage and associated new phenomena (new properties and reactions in this cage) are discussed here.

8.2 Self-assembly of Hollow Complexes

Prior to our report on the self-assembly of M6L4 cage 2, Saalfrank prepared intriguing M4L6 cages (3) aggregated from various metal ions (Mg2þ, Mn2þ, Fe2þ, Co2þ,

8.2 Self-assembly of Hollow Complexes

j279

Ni2þ, and Zn2þ) and bridging ligands with binding sites at each end [8]. In these complexes, the shape is a tetrahedron, in which metals are located at each vertex, and the six ligands are on the edges of the tetrahedron. By inserting aromatic spacers into the ligand, the volume of the cage can be expanded [8]. However, the inclusion of organic guests into these complexes has not been reported; this could be attributed to the lack of an efﬁcient driving force for inclusion of such guests in organic solvents.

M4L6 cages (4) with the same symmetry were later reported by Raymond, these being anionic and able to accommodate cationic guests quite efﬁciently [9]. The use of this complex and related cages as molecular ﬂasks has been actively developed by their group. In particular, intriguing chemical transformations within the cages have been reported [9].

280

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Representative self-assembled coordination cages reported so far (since Saalfranks 1988 report) are summarized in Table 8.1 [10–15]. An M12L24 hollow complex reported recently is one of the largest frameworks deﬁned by X-ray analysis [15g–i]. 8.2.1 M6L4 Octahedral Complex

Of the various hollow complexes, the Pd(II)-cornered M6L4 octahedral complex 2 features the following characteristics rarely shown by other complexes: 1. Quantitatively assembled from commercially available reagents: 100 g scale synthetic procedure has been established and the cage complex 2 itself is now commercially available from Wako Pure Chemicals Industries Ltd. (Pd-Nanocage). 2. Highly soluble in water because of the exposure of cationic Pd (II) units at the periphery of the complex. 3. Strongly binds large neutral organic guests through efﬁcient hydrophobic interaction in water. Because of the electrondeﬁcient nature of the ligand, charge transfer interaction also favors electron-rich guests. 4. Possesses an extraordinarily large cavity, with a diagonal Pd-Pd distance of 2.2 nm, capable of binding as many as four adamantane or carborane molecules [16]. 5. Highly symmetric, with Td symmetry, rendering the twelve pyridyl nuclei equivalent. Once guests are accommodated, however, the host symmetry is often reduced as it adapts to the included guest (or its aggregate). By analyzing the symmetry of the host, therefore, the guest geometry in the cage can be elucidated [17]. 6. Mediates photo-induced electron transfer from the guest to the host because of the extremely electron-deﬁcient triazine ring of the ligand, which is made electron-poor by three metalcoordinated pyridyl rings. Based on these characteristics, cage 2 shows unique properties not demonstrated in previously known organic hosts such as crown ethers [18], cyclodextrins [19], calixarenes [20,21], calixresorcarenes [21], cucurbituril [22], and so on. The details of these properties are discussed in the following sections. 8.2.2 Large-scale Production of the M6L4 Complex

As already mentioned, complex 2 is easily prepared on a large scale and isolated in high purity. Normally, an ethylenediamine (en) group is placed on the Pd(II) centers as the cis-protective group. This protective group is often replaced by 2,20 -bipyridine

8.2 Self-assembly of Hollow Complexes Table 8.1 Self-assembled hollow compounds composed of metals and multidentate ligands. Schematic drawings of the hollow frameworks and components (metals and ligands) are listed.

j281

282

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

8.2 Self-assembly of Hollow Complexes

j283

284

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

8.2 Self-assembly of Hollow Complexes

j285

286

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

or N,N,N0 ,N0 -tetramethylethylenediamine because the complex crystallizes more easily, which is convenient for crystallographic analysis of various host-guest complexes [23]. The metal center, Pd(II), is often replaced by Pt(II) because the Pt(II) analogs are more robust than the Pd(II)-containing cage [24]. Because of the

8.2 Self-assembly of Hollow Complexes

j287

inertness of the Pt(II)-pyridine coordination bond, however, the Pt(II) cage requires the solution to be heated for a few weeks in the presence of a template to facilitate and complete the self-assembly. Although the synthesis is more tedious, once formed, the Pt(II) cage is more stable than the Pd(II) cages and it is tolerant of a relatively large pH range [24]. All the derivatives mentioned above are structurally deﬁned by crystallographic analyses, which showed that all the cages have cavities of nearly the same shape and size, independent of the protective group (en, 2,20 -bpy, or tmen) or metal center [Pd(II) or Pt(II)].

288

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

8.3 Inclusion Properties 8.3.1 Inclusion of Adamantane and Carborane

By suspending excess adamantane (5a) in the aqueous solution of 2, the waterimmiscible 5a was gradually dissolved and included into the large hydrophobic cavity of 2, forming the 2(5a)4 inclusion complex [16]. When complex 2 was titrated with aqueous sodium adamantane carboxylate (5b), a 1:4 complex 2(5b)4 was observed (Figure. 8.1a), irrespective of the amount of the guest. For example, 2(5b)4 was formed in 25% yield, leaving 75% of empty cage 2 upon addition of one equivalent of 5b. Similarly, 2(5b)4 (50%) and the empty cage (50%) were independently observed upon the addition of two equivalents of 5b. The chemical shifts of the 1:4 complex did not change at all during the titration, and no other host-guest complexes were observed [7]. These facts indicated that the 2(5b)4 complex is much more stable than the intermediary 2(5b)n (n ¼ 1–3) complexes, and a strong positive cooperative effect is present among the guests in the formation of the 1:4 complex. The overwhelmingly dominant formation of 1:4 complexes was also observed for the reaction of 2 with 1- and 2-adamantanol [23]. Analysis of the upﬁeld shift of the individual protons of these guests elucidated the guest geometry, in which the hydrophilic OH group extends out from the cavity of the cage. This geometry was, in fact, conﬁrmed by X-ray crystallographic analysis (Figure 8.1b). o-Carborane, whose diameter (8 Å) is slightly larger than that of adamantane (7 Å), also formed

8.3 Inclusion Properties

j289

Figure 8.1 Crystal structures of (a) 2(5b)4, (b) 2(1-adamantanol)4, and (c) 2(o-carborane)4.

a similar 1:4 complex, as conﬁrmed by NMR spectroscopy and X-ray crystallographic analysis (Figure 8.1c) [23]. 8.3.2 Inclusion Geometry

Cage 2 formed 1:2 complex with a-diketone 6 (Figure 8.2) [17]. The 1 H NMR of the 1:2 complex displayed twelve signals (with the same intensity) for the triazine ligand. Associated with 2D NMR measurements, it was conﬁrmed that four triazine ligands involved in 2 were equivalent, yet the twelve protons on each ligand were all inequivalent (Figure 8.2b). This observation indicates that the ﬁxation of the included guests in the cage adopts S4 symmetry. In fact, crystallographic analysis of the 1:2 complex shows that each a-diketone 6 adopts a twisted conformation (with a dihedral angle of ca. 80 between the two carbonyl groups) and an enantiomeric pair forms a meso aggregate with S4 symmetry (Figure 8.3).

290

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.2 1 H NMR spectra of the aromatic region of (a) cage 2 and (b) 2(6)2 complex in D2O.

Figure 8.3 Crystal structure of the 2(6)2 complex.

8.3 Inclusion Properties

j291

Figure 8.4 1 H NMR spectrum of the aromatic region of cage 27 in D2O.

In the 1 H NMR spectrum of cage 2 accommodating cyclic trisilanol 7, eight signals were observed for the triazine ligand with the same integral ratio, indicating that the guest was ﬁxed along a C3 axis of the cage (Figure 8.4) [17]. The triazine ligand perpendicular to the C3 axis showed two different types of protons (PyHa and PyHb). Thus, the free rotation of the m-tolyl group of 7 on the NMR time scale affords C3v (rather than C3) symmetry. In the solid state, the m-tolyl ﬂipping is restricted and crystallographic analysis showed the inclusion of C3 symmetric guest with three m-tolyl groups arrayed in the same direction (Figure 8.5) [17]. From these spectroscopic and crystallographic observations, it can be concluded that (a) the symmetry of the guest aggregation can be elucidated by the symmetry analysis of the host rather than that of the guests, and (b) there is a good correlation between the guest geometry in solution and that in the crystalline state, probably because the atmosphere in the cage is nearly independent of whether it is in solution or the crystalline state. Thus, the cavity of the cage is essentially isolated from outside environment. 8.3.3 Bimolecular Recognition

When two different guests (Ga and Gb) are added to the solution of cage 2, a heterotopic inclusion complex 2GaGb can be more stable than homotopic

292

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.5 Crystal structure of cyclic trisilanol 7 within cage 2.

complexes 2(Ga)2 or 2(Gb)2. If the ﬁrst guest Ga is a planar aromatic compound, it stacks with a ligand of the cage, leaving a roughly spherical void in the cavity. Thus, a spherical guest Gb ﬁts well. An interesting example of this is the pairwise selective recognition of perylene (8) and cis-decalin (9) [25]. These guests were unable to form homotopic inclusion complexes [2(8)2 or 2(9)2] when they were individually combined with cage 2. Nevertheless, when they were combined together and suspended in the aqueous solution of 2, a stable heterotopic complex 289 was formed in good yield, despite the absence of any special attractive forces between the guests 8 and 9. Molecular modeling displayed a stable guest packing with C3v symmetry (assuming free rotation of the guests along the C3 axis), where the perylene molecule stacked on the triazine ligand and the remaining void is occupied by cis-decalin (Figure 8.6). Although there are many reports on similar bimolecular recognition (or ternary complex formation), in most cases, attractive forces such as charge-transfer interactions are present between the two guests. In contrast, two guests have to be spatially paired in the cavity of 2 despite the absence of any strong afﬁnity between the guests. In addition to the combination of perylene and cis-decalin, highly pairwise selectivities were observed for various combinations of two different guests (e.g., pyrene and cis-decalin, or pyrene and hexane), often by adjusting the steric demand of one substrate. The facile formation of the ternary complexes (2GaGb) is an important issue for designing cavity-directed chemical reactions between two substrates, as discussed below.

8.3 Inclusion Properties

j293

Figure 8.6 Molecular modeling of heterotopic inclusion complex 289.

8.3.4 Recognition of Bulky Guests

Cage 2 possesses four portals whose dimensions are considerably smaller than that of the cages cavity itself. Since Pd-N dissociation is a very slow process in comparison with the host-guest complexation, guest molecules should enter the cavity without breaking any bonds of the cage. Therefore, a guest which is larger than the host portal is expected to enter by expanding the portal framework with some energy barrier. A bulky guest, 1,3,5-tri(t-butyl)benzene (10), is not included in the cage at room temperature because of the high energy barrier. On heating the solution to 80 C, however, this bulky guest was included in ca. 45% yield into the cavity via thermal slippage [23]. After the solution was cooled down to room temperature, the guest (10) was no longer able to escape from the cavity. In this way, the large guest was incarcerated (Figure 8.7). Tetrabenzylsilane (11) has a number of conformational isomers in equilibrium. Interestingly, during the inclusion process of this large guest, symmetry-broken spectra were observed for intermediary species before obtaining a simple 1 H NMR spectrum, indicating the formation of the highly symmetric 211 complex, whose structure was conﬁrmed by X-ray analysis (Figure 8.8) [23]. It is therefore reasonable

294

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.7 Crystal structure of 1,3,5-tri(t-butyl)benzene (10) within cage 2.

Figure 8.8 Crystal structure of tetrabenzylsilane (11) within cage 2.

8.3 Inclusion Properties

j295

that structures including less than four benzyl groups of 11 exist as intermediates before the guest is ﬁnally trapped in the cavity of 2. 8.3.5 The Recognition of Azobenzene and Stilbene

Since the cavity of cage 2 is surrounded by four ligands arrayed in S4 symmetry, guest molecules or aggregates with S4 symmetry interact efﬁciently with the cage. In the complexation with 4,40 -dimethylstilbene, the cis isomer was bound quite strongly to give quantitatively a 1:2 inclusion complex, whereas the trans isomer showed no interaction [26]. The preferential complexation of the cis isomer is explained by assuming the formation of a tennis ball-like aggregation with an S4 symmetry from two cis-4,40 -dimethylstilbene molecules. In fact, this proposed aggregation was strongly suggested by molecular modeling study (Figure 8.9). A similar cis-selective inclusion was observed for 4,40 -dimethylazobenzene. Although the cis-trans isomerization of azobenzene by UV irradiation is a wellknown process, 100% conversion from trans to cis at a photostationary state is normally impossible. In the presence of cage 2, however, only the cis isomer is trapped by the cage via the formation of the stable tennis-ball like S4 dimer, pushing the cis-trans photo-equilibration overwhelmingly toward the cis-isomer [26].

Figure 8.9 Optimized structure for the tennis ball-like aggregation of two 4,40 -dimethylstilbene molecules within 2.

296

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

8.3.6 Molecular Ice

Among various potential guests for cage 2, the simplest is a water molecule. The cavity of the cage is expected to be ﬁlled with water in aqueous solution when organic guests are absent. We wondered whether the included water molecules would behave as bulk water or adopt a more highly ordered state. On the surface of hydrophobic substances, water molecules are known to develop an ordered hydrogen-bonded network to neutralize their own high polarity. The origin of this hydrophobic interaction is considered to be the entropic advantage when the high-potential water molecules at the surface become free from the restrictions of the network upon contact with the hydrophobic species. Thus, in the hydrophobic cavity of the cage, endohedral clusterization (network formation) of water is expected. Crystallographic analysis was therefore carried out for the empty cage that had been crystallized from water. Interestingly, ten water molecules form an adamantane-like (H2O)10 cluster in the cavity of cage 2, as expected (Figure 8.10) [27]. This cluster structure is the smallest unit of the 3D hydrogen-bonded network of naturally occurring Ic-type ice. If the ice structure is deﬁned by the 3D network of water molecules, the (H2O)10 cluster in our cavity is regarded as molecular ice. The

Figure 8.10 Crystal structure of an adamantane-like (H2O)10 cluster within cage 2.

8.3 Inclusion Properties

j297

Figure 8.11 (a) Neutron diffraction analysis of a (D2O)10 cluster within cage 2, and (b) the corresponding structural drawing.

crystallographic analysis showed that the cavity size is ideal for binding the molecular ice. As shown in Figure 8.10, the adamantane-like framework of the molecular ice has almost no distortion in the framework as compared with the unit of Ic-type ice; the molecular ice is only slightly compressed by the four ligands of the cage, making O s s s O s s O angles (108.8–122.3 ) at the bridgehead of the adamantane framework slightly larger than that of Ic-type ice (109.5 ). A question arises as to whether this molecular ice cluster is recognized via OH s s s p hydrogen bonding or HO: s s s p (n-p) interaction in the cage. The answer is not to be derived from X-ray crystallography, because hydrogen atoms were not located in the crystallographic analysis. Therefore, neutron diffraction was used to distinguish whether or not a hydrogen atom exists between the bridgehead water oxygen and the ligand. Single crystals of the cage were grown from D2O solution. The neutron diffraction analysis of the molecular ice structure (D2O)10 in the cage clearly revealed the absence of deuterium nuclear density between the bridgehead oxygen and the ligand (Figure 8.11) [27]. This observation indicates that the bridgehead water is recognized by the ligands via HO: s s s p (n-p) interaction not OH s s s p hydrogen bonding. In contrast to a typical OH s s s p interaction, where an aromatic ring is an electron donor and hydrogen is an acceptor, the extremely electron-poor triazine ring seems to work as an electron acceptor to the lone pair of the oxygen of water. This reverse charge donation is probably the driving force for the formation of a stable molecular ice cluster. 8.3.7 Peptide Recognition

Recognition of a speciﬁc amino acid sequence from a protein or a speciﬁc tag peptide sequence is expected to lead to variety of medical and materials applications. Whereas the recognition of a single amino acid residue by synthetic hosts has been reported previously, the artiﬁcial recognition of a sequence of two or more amino acid residues has been quite limited. There are only a few reports on the

298

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.12 Crystal structure of cage 2 accommodating Ac-Trp-Trp-Ala-NH2 (12).

sequence-selective recognition of peptides by synthetic hosts, where two amino acid receptors are simply connected to recognize dipeptides. Cage 2 is able to recognize an oligopeptide sequence because its binding pocket can accommodate up to three amino acid residues. Therefore, various oligopeptides were prepared and the interaction with 2 examined. The X-X-Ala tripeptides (where X denotes aromatic residue such as Phe, Tyr, Trp) were sequence-selectively bound by the cage [28]. The largest association constant was observed for Ac-Trp-Trp-Ala-NH2 (12; Ka > 106 M1), whose geometry in the cage was crystallographically analyzed (Figure 8.12). The crystal structure revealed that the indole rings of two Trp groups and the methyl group of Ala efﬁciently interacted with the cage via p-p and CH-p interactions, respectively. Even the Ala-Gly mutant, Ac-Trp-Trp-Gly-NH2, showed a considerably smaller association constant (Ka ¼ 7.4 · 104 M1), indicating the importance of cooperativity of the above three interactions. The sequence of the three residues was particularly important: Ac-Trp-Ala-Trp-NH2 and Ac-Ala-Trp-TrpNH2, which have the same residues but different sequence, were poorly recognized. These results represent the ﬁrst example of sequence-selective recognition of peptides by a single recognition site. The crystal structure of 212 also revealed that the N-terminal of 12 contacts with the p-face of the ligand, whereas the C-terminal is directed toward the portal of the cage. Hence, as expected, oligopeptides extending the sequence at the C-terminal were also bound strongly, whereas those extending the sequence at the N-terminal were not.

8.4 New Physical Properties

j299

8.4 New Physical Properties

Since the environment of the cavity presented by 2 is spatially and electronically conﬁned, physical properties of encapsulated molecules are expected to change to some extent, or even dramatically. In this section, the variable properties of guests in the cavity of cage 2 are discussed. 8.4.1 Redox Control of Ferrocene

The formation of a 1:4 complex 2(13)4 (13: ferrocene) was conﬁrmed both by NMR spectroscopy and crystallographic analysis (Figure 8.13) [29]. The oxidation potential of ferrocene to ferrocenium ion [Fe(III)] in the cage was anodically shifted by 73 mV as compared to that in water. Thus, the cationic cage 2 suppresses the oxidation of ferrocene to ferrocenium ion. 8.4.2 Induction of Intermolecular Spin—spin Interaction

Cage 2 can bind two organic radical molecules such as 2-naphthyl nitronyl nitroxide (14). Normally, the organic spin centers are protected by bulky substituents to enhance

Figure 8.13 Crystal structure of 2(13)4 complex.

300

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

the kinetic stability of the radical. Because of the steric demand of the bulky substituents, intermolecular spin-spin interaction can be difﬁcult to observe for common organic radicals. As for the radical 14, the intermolecular spin-spin interaction is not observed even in the solid state because of the offset packing of the molecules, which distances the radical centers from each other. However, when a solution of the 1:2 inclusion complex 2(14)2, obtained by suspending solid 14 in the aqueous solution of 2, was frozen at low temperatures and subjected to ESR study, new broad signals ascribable to the intermolecular spin-spin interaction of the included guests were clearly observed (Figure 8.14) [30]. The spin-spin interaction was unambiguously conﬁrmed by the observation of Dms ¼ 2 signal, which is attributed to the triplet state

Figure 8.14 ESR spectra of (a) 14 at 293 K and 2(14)2 at (b) 363, (c) 293, (d) 273, and (e) 103 K.

8.5 New Reactions

j301

Figure 8.15 Crystal structure of 2(14)2 complex.

Q1

of two interacting spins, at half the resonance magnetic ﬁeld. The crystal structure (Figure 8.15) also conﬁrmed the proximity of the two spin centers. The averaged spinspin distance is ca. 5.8 Å, consistent with those estimated from the ESR spectrum according to the equation: D0 ¼ 3/2(m0gb)/(4pr3). For a series of organic radicals 14–17, the interaction decrease in the order 14 > 15 > 16 > 17 [31]. Consistent with this observation, crystallographic studies revealed that the spin-spin distances increased in the same order.

8.5 New Reactions

The cavity of cage 2 can normally accommodate two or more guest molecules. The proximity of reaction centers of two substrates conﬁned in the cavity of the cage

302

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

reduces the entropy cost signiﬁcantly and controls the orientation of the substrates, leading to a remarkable enhancement of reaction rates as well as the induction of high stereo- and regioselectivities. In addition, quite unusual reactions have been found because of the high binding ability and electron-deﬁcient nature of the cage. Intriguing and sometimes unexpected reactions occurring in the cage structure of 2 are discussed in the following sections. 8.5.1 [2þ2] Oleﬁn Photodimerization

Upon irradiation of the cage accommodating two oleﬁnic compounds, [2þ2] photodimerization often takes place quite efﬁciently with high stereoselectivity. A wellstudied substrate for the [2þ2] photoaddition is acenaphthylene (18), which is known to react at the 1,2 position to give a mixture of syn- and anti-isomers, wherein the yields and the stereoselectvity are not excellent. This reaction was examined in cage 2. First, we observed the quantitative formation of 1:2 host-guest complex [2(18)2] by suspending powdered 18 in an aqueous solution of 2. Upon irradiation with a high-pressure mercury lamp, the reaction was complete within 30 min to give the syn [2þ2] adduct 19 stereoselectively (Figure 8.16) [32]. In a similar way, the [2þ2] photoaddition of 1-methylacenaphthylene was examined. In common organic solvents, this substrate does not afford the [2þ2] dimer because of the steric demands of the methyl group at the 1-position. In cage 2, however, the reaction took place smoothly to give only the syn and head-to-tail isomer (Figure 8.16) [32]. Thus, both stereo- and regioselectivies were perfectly controlled in the reaction of an otherwise unreactive substrate.

Figure 8.16 [2+2] Photodimerization of acenaphthylene (18) and 1-methylacenaphthylene within cage 2.

8.5 New Reactions

j303

Figure 8.17 Pairwise-selective [2 + 2] photodimerization of acenaphthylene (18) and 5-ethoxy-1,4-naphthoquinone (20) within cage 2.

8.5.2 Pairwise-selective Oleﬁn Photodimerization

In common organic solvents, irradiation of a mixture of two oleﬁns provides a mixture of homo- and hetero-dimers (and their regio- and stereoisomers) unless one of the substrates is used in large excess. By utilizing the large cavity of cage 2, a cross dimer is often selectively formed from two different oleﬁns. For example, by suspending two oleﬁns 18 and 5-ethoxy-1,4-naphthoquinone (20) in an aqueous solution of 2, the ternary complexes 21820 was quantitatively formed. On irradiation, cross adduct 21 was formed in 92% yield with 100% syn stereoselectivity (Figure 8.17) [33]. 8.5.3 Unusual [2þ2] Photoaddition

In many cases, unusual [2þ2] photoadditions have been observed, reactions that typically would never take place in common solvents under ordinary photoreaction conditions. Pyrene (22) was reacted with N-cyclohexylmaleimide (23) after forming a ternary complex 22223 and being irradiated for 6 h to give the cross [2þ2] adduct 24 in 77% yield based on 2 (Figure 8.18) [34]. The [2þ2] photoaddition of maleimide 23 with ﬂuoranthene (25), whose photoaddition has never been reported, is also intriguing. The LUMO orbital of 25 is not localized at a speciﬁc C¼C bond, but rather equally delocalized over the molecule, making the reaction site unpredictable. When ternary complex 22325 was irradiated (365 nm) in D2O for 5 min at room temperature, a single product 26 was obtained that arises from via [2þ2] photoaddition at the C2¼C3 site of 25 (Figure 8.18) [34]. The high regioselectivity is presumably ascribed to the geometric control of the inclusion complex deﬁned by the conﬁned cavity of 2. 8.5.4 Diels-Alder Reaction

In addition to [2þ2] photoadditions, thermal [4þ2] cycloadditions (Diels-Alder reactions) are also remarkably accelerated in the cavity of cage 2. The reaction of

304

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.18 Unusual [2+2] photoadditions of maleimide (23) and pyrene (22) or ﬂuoranthene (25) within cage 2.

benzoquinone and 2,3-dimethyl-1,3-butadiene in the cage proceeds approximately 100 times faster than the analogous reaction in a common organic solvent or an aqueous suspension [35]. Surprisingly, some aromatic compounds which otherwise show no Diels-Alder reactivity participate in Diels-Alder reactions when they are coenclathrated with a dienophile in the cage. For example, triphenylene (27) and maleimide 23 were pairwise selectively bound in the cavity in a moderate yield. Upon heating the solution, these two substrates underwent a Diels-Alder cycloaddition in which one of the benzene rings of 27 participated as a diene [34]. NMR spectroscopic analysis showed that the included triphenylene was quantitatively converted into the unusual adduct 28, in an exo-selective fashion, without the formation of any by-products (Figure 8.19). In the reaction with perylene (8), the Diels-Alder adduct 29 was also formed in a high yield. Although this reaction proceeds to some extent in common organic solvents, the adduct is easily airoxidized to reproduce the perylene aromaticity. In the cage, however, the adduct was protected from oxidation and the adduct 29 was observed as a stable product within 2 (Figure 8.19) [34]. Anthracene (30), a classical substrate for the Diels-Alder reaction, always reacts with dienophile at its 9,10-position because of the highest HOMO level at this position. To our surprise, the reaction of 30 with dienophile 23 proceeded at a terminal, rather than a central, benzene ring of 30 to give unusual adduct 31 with exo-selectivity (Figures 8.20 and 8.21) [36]. Molecular modeling predicted that, because of the steric demand of the N-cyclohexyl group, dienophile 23 is only allowed access to the terminal benzene ring, precluding the reaction at the central ring (Figure 8.21b).

8.5 New Reactions

Figure 8.19 Diels-Alder reactions of maleimide (23) and triphenylene (27) or perylene (8) within cage 2.

Figure 8.20 Diels-Alder reactions of maleimide (23) and anthracene (30) within cage 2.

Figure 8.21 (a) Crystal structure of 231 complex (after Diels-Alder reaction) and (b) optimized structure of ternary complex 22330 (before the reaction).

j305

306

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

8.5.5 Alkane Oxidation

In the above photo and thermal cycloaddition reactions, the reaction pathway is restricted by the space of the cavity reducing the energy barrier of a speciﬁc pathway while enhancing the others. The cage itself, however, does not function electronically as a catalytic redox medium. In contrast, an alkane photo-oxidation takes place in the cage via the cage-mediated electron transfer. When the aqueous solution of 1:4 adamantane inclusion complex 2(5a)4 was irradiated (>300 nm) under anaerobic conditions, the solution color turned blue, indicating the generation of radical species. Analysis of the substrate after irradiation indicated the formation of adamantan-1-ol (as a mixture with the corresponding peroxide) in 24% yield, suggesting that one out of four included guest molecules was oxidized (Figure 8.22) [37]. Electrochemical measurement showed a ﬁrst one-electron reduction potential of the cage that was considerably low compared with the subsequent second through fourth reduction potentials. From these facts, the most probable pathway to promote the alkane oxidation includes (a) excitation of the triazine ligand with a low-lying LUMO, followed by (b) a one-electron transfer from the HC bond at the bridgehead of adamantane to the ligand, (c) the resulting adamantane radical cation is immediately dissociated into adamantyl radical and proton, and (d) the adamantane radical is then immediately trapped by dissolved molecular oxygen or water (which was conﬁrmed by experiments using 18 O2 and H2 18 O). Since the cage can accept only one electron, as suggested by electrochemical measurements, only one adamantane guest is oxidized (in 96% yield based on cage 2). The crystal structure of 2 (5a)4 shows that the adamantane CH bond and the triazine ligand are within van der Waals contact, presumably facilitating the unusual photo-induced alkane oxidation mediated by the electron-deﬁcient cage [37]. Similar oxidations were also promoted with cycloalkanes to give the corresponding alcohols and ketones. 8.5.6 Wacker Oxidation

The Pd(II) hinge in cage 2 can also participate in a chemical transformation. The catalytic Wacker-type oxidation of oleﬁns took place when 8-nonen-1-ol (32) was heated for 5 h at 80 C in the presence of cage 2 (5 mol%), giving 9-hydroxynonan-2-one (33)

Figure 8.22 Photo-oxidation of adamantane (5a) within cage 2 via guest-host electron transfer.

8.5 New Reactions

j307

Figure 8.23 Catalytic Wacker oxidation of 8-nonen-1-ol (32) and styrene in the presence of cage 2.

in 66% yield (Figure 8.23) [38]. The crystal structure of 1:2 nonan-1-ol (34) inclusion complex [2(34)2] revealed the proximity of the terminal carbon of 34 to the Pd(II) center (Figure 8.24), implying a similar placement of the oleﬁn part of 32 with respect to the Pd(II) center, as found in the [2(32)2] complex. The Pd(0) species that would result from the Wacker-type oxidation is reoxidized into Pd(II) despite the absence of

Figure 8.24 Crystal structure of 2(34)2 complex.

308

j 8 New Properties and Reactions in Self-assembled M L

6 4

Coordination Cages

Figure 8.25 Sol-gel condensation of PhSi(OMe)3 (35) within case 2 to give cyclic trimer 36.

any reoxidant. It is likely that dissolved molecular oxygen in the aqueous solution is sufﬁcient to promote the reoxidation of the Pd(0) species. Similar Wacker-type oxidation was also observed for other oleﬁns such as styrene, but a slight excess of (en)Pd (NO3)2, the hinge component of cage 2, is necessary to attain formation of acetophenone in high yield (Figure 8.23) [39]. A two-phase reaction can be proposed for the catalytic cycle. First, styrene (organic phase) is dissolved into the cavity of 2 in the aqueous phase, where the oxidation is promoted. Then, the more hydrophilic product, acetophenone, is replaced by unreacted styrene to promote the catalytic cycle. 8.5.7 Discrete Siloxane Synthesis

While all the reactions discussed above are organic transformations, we also examined inorganic transformations in the cage. Inorganic reactions often provide structurally inﬁnite bulk materials. For example, hydrolysis and dehydration of trialkoxysilanes, well-known as sol-gel condensation, gives a three-dimensional SiO network, providing an efﬁcient low-temperature synthesis of functionalized silica gel. When the sol-gel condensation of PhSi(OMe)3 (35) was carried out in the presence of cage 2, the selective formation of the all-cis isomer of a cyclic trimer 36 was observed (Figure 8.25) [40,41]. In the common solution-state reaction, it is known that a stereoisomeric mixture of cyclic tetramers is predominantly formed, which is subsequently condensed into a 3D network because of the presence of reactive SiOH groups. In the cavity of 2, the cyclic trimer 36 is stabilized, and, despite the presence of the reactive SiOH group, it is unreactive toward further polymerization even after several weeks in solution.

8.6 Conclusion

New chemistry – particularly, new reactions and properties – created by exploiting the conﬁned cavity of self-assembled M6L4 coordination cages 2 is discussed in this chapter. Thanks to the large cavity and strong binding ability of the cage, as compared with many other known host compounds, this cage allows us to examine binding and chemical transformations that have previously been impossible. This

References

j309

provides for the continual discovery of new reactions and properties within the conﬁned cavity of cage 2 since its ﬁrst synthesis in 1995. The chemistry of space promises creative research and will last forever. It is therefore our hope that this chapter will be welcomed by readers as an inspiring source for creative research.

8.7 Experimental: Synthesis of M6L4 Cage 2

In a 100-mL round-bottom ﬂask, (en)Pd(NO3)2 (2.00 g, 6.88 mmol) and 2,4,6-tris(4-pyridyl)-1,3,5-triazine (1.51 g, 4.83 mmol) were combined with H2O (60 mL). When the suspended mixture was stirred at 100 C for 3 h, a pale yellow solution was obtained. The resulting solution was ﬁltered to remove excess triazine ligand. After removal of the solvent by evaporation under reduced pressure, 2 was obtained as a pale yellow powder in 93% yield (3.20 g, 1.07 mmol). 1 H NMR (500 MHz, D2O, TMS as an external standard): d 2.98 ppm (s, 24H, -CH2-), 8.63 ppm (dd, J ¼ 6.8, 3.9 Mz, 24H), 9.13 ppm (dd, J ¼ 6.8, 3.9 Hz, 24H). 13 C NMR (125 MHz, D2O, CHCl3 as an external standard): d 46.9 (CH), 125.5 (CH), 145.6 (Cq), 152.2 (CH), 169.8 (Cq).

Acknowledgments

The authors would like to express sincere thanks to many co-workers who, with their efforts and creative minds, have signiﬁcantly contributed to the development of the chemistry discussed in this article. Their names are found in many of the following references.

Abbreviations

en bpy tmen Py

ethylenediamine 2,20 -bipyridine N,N,N0 ,N0 -tetramethylethylenediamine pyridyl

References 1 Fujita, M., Yazaki, J. and Ogura, K. (1990) J. Am. Chem. Soc., 112, 5645–5647. 2 Stang, P.J. and Cao, D.H. (1994) J. Am. Chem. Soc., 116, 4981–4982. 3 (a) Lahav, M., Gabai, R., Shipway, A.N. and Willner, I. (1999) Chem.

Commun., 1937–1938. (b) Bélanger, S., Hupp, J.T., Stern, C.L., Slone, R.V., Watson, D.F. and Carrell, T.G. (1999) J. Am. Chem. Soc., 121, 557–563. (c) Lau, V.C., Berben, L.A. and Long, J.R. (2002) J. Am. Chem. Soc., 124, 9042–9043.

310

j 8 New Properties and Reactions in Self-assembled M L

6 4

4 (a) Drain, C.M. and Lehn, J.-M. (1994) J. Chem. Soc. Chem. Commun., 2313– 2315. (b) Stang, P.J., Fan, J. and Olenyuk, B. (1997) Chem. Commun., 1453–1454. 5 Olenyuk, B., Whiteford, J.A. and Stang, P.J. (1996) J. Am. Chem. Soc., 118, 8221–8230. 6 (a) Stang, P.J., Olenyuk, B., Fan, J. and Arif, A.M. (1996) Organometallics, 15, 904–908. (b) Stang, P.J., Cao, D.H., Chen, K., Gray, G.M., Muddiman, D.C. and Smith, R.D. (1997) J. Am. Chem. Soc., 119, 5163–5168. 7 Fujita, M., Oguro, D., Miyazawa, M., Oka, H., Yamaguchi, K. and Ogura, K. (1995) Nature, 378, 469–471. 8 (a) Saalfrank, R.W., Stark, A., Peters, K. and von Schnering, H.G. (1988) Angew. Chem. Int. Ed. Engl., 27, 851–853; Angew. Chem., 100, 878–880., (b) Saalfrank, R. W., Stark, A., Bremer, M. and Hummel, H.-U. (1990) Angew. Chem. Int. Ed. Engl., 29, 311–314; Angew. Chem., 102, 292–295., (c) Saalfrank, R.W., Hörner, B., Stalke, D. and Salbeck, J. (1993) Angew. Chem. Int. Ed. Engl., 32, 1179–1182; Angew. Chem., 105, 1223–1225. 9 (a) Beissel, T., Powers, R.E. and Raymond, K.N. (1996) Angew. Chem. Int. Ed. Engl., 35, 1084–1086; Angew. Chem., 108, 1166–1168. (b) Caulder, D.L., Powers, R.E., Parac, T.N. and Raymond, K.N. (1998) Angew. Chem. Int. Ed., 37, 1840–1843; Angew. Chem., 110, 1940–1943. (c) Ziegler, M., Brumaghim, J.L. and Raymond, K.N. (2000) Angew. Chem. Int. Ed., 39, 4119–4121; Angew. Chem., 112, 4285–4287. (d) Fiedler, D., Bergman, R.G. and Raymond, K.N. (2004) Angew. Chem. Int. Ed., 43, 6748–6751; Angew. Chem., 116, 6916–6919. (e) Dong, V.M., Fiedler, D., Carl, B., Bergman, R.G. and Raymond, K.N. (2006) J. Am. Chem. Soc., 128, 14464–14465. 10 (a) Mann, S., Huttner, G., Zsolnai, L. and Heinze, K. (1996) Angew. Chem. Int. Ed.

Coordination Cages

Engl., 35, 2808–2809; Angew. Chem., 108, 2983–2984. (b) Fleming, J.S., Mann, K.L. V., Carraz, C.-A., Psillakis, E., Jeffrey, J. C., McCleverty, J.A. and Ward, M.D. (1998) Angew. Chem. Int. Ed., 37, 1279–1281; Angew. Chem., 110, 1315–1318. (c) Amoroso, A.J., Jeffery, J. C., Jones, P.L., McCleverty, J.A., Thornton, P. and Ward, M.D. (1995) Angew. Chem. Int. Ed. Engl., 34, 1443–1446; Angew. Chem., 107, 1577–1580. (d) Brückner, C., Powers, R. E. and Raymond, K.N. (1998) Angew. Chem. Int. Ed., 37, 1837–1839; Angew. Chem., 110, 1937–1940. (e) Umemoto, K., Yamaguchi, K. and Fujita, M. (2000) J. Am. Chem. Soc., 122, 7150–7151. (f) Saalfrank, R.W., Glaser, H., Demleitner, B., Hampel, F., Chowdhry, M.M., Schünemann, V., Trautwein, A.X., Vaughan, G.B.M., Yeh, R., Davis, A.V. and Raymond, K.N. (2002) Chem. Eur. J., 8, 493–497.(g) Schweiger, M., Seidel, S. R., Schmitz, M. and Stang, P.J. (2000) Org. Lett., 2, 1255–1257. (h) Bourgeois, J.-P., Fujita, M., Kawano, M., Sakamoto, S. and Yamaguchi, K. (2003) J. Am. Chem. Soc., 125, 9260–9261. (i) Hiraoka, S., Yi, T., Shiro, M. and Shionoya, M. (2002) J. Am. Chem. Soc., 124, 14510– 14511. (j) Hiraoka, S., Harano, K., Shiro, M. and Shionoya, M. (2005) Angew. Chem. Int. Ed., 44, 2727–2731; Angew. Chem., 117, 2787–2791. (k) Albrecht, M., Janser, I., Meyer, S., Weis, P. and Fröhlich, R. (2003) Chem. Commun., 2854–2855. 11 (a) Takeda, N., Umemoto, K., Yamaguchi, K. and Fujita, M. (1999) Nature, 398, 794–796. (b) Umemoto, K., Tsukui, H., Kusukawa, T., Biradha, K. and Fujita, M. (2001) Angew. Chem. Int. Ed., 40, 2620–2622; Angew. Chem., 113, 2690–2692. (c) Ikeda, A., Yoshimura, M., Udzu, H., Fukuhara, C. and Shinkai, S. (1999) J. Am. Chem. Soc., 121, 4296– 4297. (d) Müller, I.M. and Möller, D. (2005) Angew. Chem. Int. Ed., 44, 2969– 2973; Angew. Chem., 117, 3029–3033.

References

(e) Fujita, M., Nagao, S. and Ogura, K. (1995) J. Am. Chem. Soc., 117, 1649– 1650. (f) Hiraoka, S. and Fujita, M. (1999) J. Am. Chem. Soc., 121, 10239–10240. (g) Sun, X., Johnson, D. W., Caulder, D.L., Powers, R.E., Raymond, K.N. and Wong, E.H. (1999) Angew. Chem. Int. Ed., 38, 1303–1307; Angew. Chem., 111, 1386–1390. (h) Catalano, V.J., Bennett, B.L., Kar, H.M. and Noll, B.C. (1999) J. Am. Chem. Soc., 121, 10235–10236. (i) Claessens, C.G. and Torres, T. (2002) J. Am. Chem. Soc., 124, 14522–14523. (j) Su, C.-Y., Cai, Y.P., Chen, C.-L., Lissner, F., Kang, B.-S. and Kaim, W. (2002) Angew. Chem. Int. Ed., 41, 3371–3375; Angew. Chem., 114, 3519–3523. (k) Su, C.-Y., Cai, Y.-P., Chen, C.-L., Smith, M.D., Kaim, W. and zur Loye, H.-C. (2003) J. Am. Chem. Soc., 125, 8595–8613. (l) Kumazawa, K., Yamanoi, Y., Yoshizawa, M., Kusukawa, T. and Fujita, M. (2004) Angew. Chem. Int. Ed., 43, 5936–5940; Angew. Chem., 116, 6062–6066. 12 (a) Jacopozzi, P. and Dalcanale, E. (1997) Angew. Chem. Int. Ed. Engl., 36, 613–615; Angew. Chem., 109, 665–667. (b) Kobayashi, K., Yamada, Y., Yamanaka, M., Sei, Y. and Yamaguchi, K. (2004) J. Am. Chem. Soc., 126, 13896–13897. (c) Stang, P.J., Olenyuk, B., Muddiman, D.C. and Smith, R.D. (1997) Organometallics, 16, 3094–3096. (d) Schweiger, M., Yamamoto, T., Stang, P. J., Bläser, D. and Boese, R. (2005) J. Org. Chem., 70, 4861–4864. (e) Hartshorn, C. M. and Steel, P.J. (1997) Chem. Commun., 541–542. (f) Liu, H.-K. and Tong, X. (2002) Chem. Commun., 1316– 1317. (g) Fox, O.D., Dalley, N.K. and Harrison, R.G. (1998) J. Am. Chem. Soc., 120, 7111–7112. (h) James, S.L., Mingos, D.M.P., White, A.J.P. and Williams, D.J. (1998) Chem. Commun., 2323–2324. (i) Cotton, F.A., Daniels, L.M., Lin, C. and Murillo, C.A. (1999) Chem. Commun., 841–842. (j) Ikeda, A., Ayabe, M., Shinkai, S., Sakamoto, S. and

j311

Yamaguchi, K. (2000) Org. Lett., 2, 3707–3710. (k) Hong, M., Zhao, Y., Su, W., Cao, R., Fujita, M., Zhou, Z. and Chan, A.S.C. (2000) J. Am. Chem. Soc., 122, 4819–4820. (l) Zhong, Z., Ikeda, A., Ayabe, M., Shinkai, S., Sakamoto, S. and Yamaguchi, K. (2001) J. Org. Chem., 66, 1002–1008. (m) Chand, D.K., Biradha, K., Fujita, M., Sakamoto, S. and Yamaguchi, K. (2002) Chem. Commun., 2486–2487. (n) Müller, I.M., Spillmann, S., Franck, H. and Pietschnig, R. (2004) Chem. Eur. J., 10, 2207–2213. (o) McKinlay, R.M., Thallapally, P.K., Cave, G.W.V. and Atwood, J.L. (2005) Angew. Chem. Int. Ed., 44, 5733–5736; Angew. Chem., 117, 5879–5882. (p) Haino, T., Kobayashi, M., Chikaraishi, M. and Fukazawa, Y. (2005) Chem. Commun., 2321–2323. (q) Hiraoka, S., Harano, K., Shiro, M., Ozawa, Y., Yasuda, N., Toriumi, K. and Shionoya, M. (2006) Angew. Chem. Int. Ed., 45, 6488–6491; Angew. Chem., 118, 6638– 6641. 13 (a) McMorran, D.A. and Steel, P.J. (1998) Angew. Chem. Int. Ed., 37, 3295–3297; Angew. Chem., 110, 3495–3497. (b) Chand, D.K., Biradha, K. and Fujita, M. (2001) Chem. Commun., 1652–1653. (c) Yue, N.L.S., Eisler, D.J., Jennings, M.C. and Puddephatt, R.J. (2004) Inorg. Chem., 43, 7671–7681. (d) Fujita, M., Yu, S.-Y., Kusukawa, T., Funaki, H., Ogura, K. and Yamaguchi, K. (1998) Angew. Chem. Int. Ed., 37, 2082–2085; Angew. Chem., 110, 2192– 2196. (e) Schnebeck, R.-D., Freisinger, E. and Lippert, B. (1999) Angew. Chem. Int. Ed., 38, 168–171; Angew. Chem., 111, 235–238. (f) Yu, S.-Y., Huang, H., Liu, H.-B., Chen, Z.-N., Zhang, R. and Fujita, M. (2003) Angew. Chem. Int. Ed., 42, 686–690; Angew. Chem., 115, 710– 714. (g) Baxter, P.N.W., Lehn, J.-M., Kneisel, B.O., Baum, G. and Fenske, D. (1999) Chem. Eur. J., 5, 113–120. (h) Garcia, A.M., Bassani, D.M., Lehn, J.M., Baum, G. and Fenske, D. (1999)

312

j 8 New Properties and Reactions in Self-assembled M L

6 4

Chem. Eur. J., 5, 1234–1238. (i) Fujita, N., Biradha, K., Fujita, M., Sakamoto, S. and Yamaguchi, K. (2001) Angew. Chem. Int. Ed., 40, 1718–1721; Angew. Chem., 113, 1768–1771. (j) Benkstein, K.D. and Hupp, J.T. (2000) Mol. Cryst. Liq. Cryst., 342, 151–158. (k) Sun, S.-S. and Lees, A.J. (2001) Chem. Commun., 103–104. (l) Sommer, R.D., Rheingold, A.L., Goshe, A.J. and Bosnich, B. (2001) J. Am. Chem. Soc., 123, 3940–3952. (m) Kuehl, C.J., Yamamoto, T., Seidel, S.R. and Stang, P. J. (2002) Org. Lett., 4, 913–915. (n) Kumazawa, K., Biradha, K., Kusukawa, T., Okano, T. and Fujita, M. (2003) Angew. Chem. Int. Ed., 42, 3909–3913; Angew. Chem., 115, 4039–4043. (o) Yoshizawa, M., Nakagawa, J., Kumazawa, K., Nagao, M., Kawano, M., Ozeki, T. and Fujita, M. (2005) Angew. Chem. Int. Ed., 44, 1810–1813; Angew. Chem., 117, 1844–1847. (p) Yoshizawa, M., Nagao, M., Kumazawa, K. and Fujita, M. (2005) J. Organomet. Chem., 690, 5383–5388. (q) Yoshizawa, M., Ono, K., Kumazawa, K., Kato, T. and Fujita, M. (2005) J. Am. Chem. Soc., 127, 10800–10801. (r) Yoshizawa, M., Kumazawa, K. and Fujita, M. (2005) J. Am. Chem. Soc., 127, 13456–13457. (s) Mahmoudkhani, A.H., Côté, A.P. and Shimizu, G.K.H. (2004) Chem. Commun., 2678–2679. 14 (a) Roche, S., Haslam, C., Adams, H., Heath, S.L. and Thomas, J.A. (1998) Chem. Commun., 1681–1682. (b) Heinrich, J.L., Berseth, P.A. and Long, J.R. (1998) Chem. Commun., 1231– 1232. (c) Klausmeyer, K.K., Rauchfuss, T. B. and Wilson, S.R. (1998) Angew. Chem. Int. Ed., 37, 1694–1696; (1998) Angew. Chem., 110, 1808–1810. (d) Klausmeyer, K.K., Wilson, S.R. and Rauchfuss, T.B. (1999) J. Am. Chem. Soc., 121, 2705– 2711. (e) Johannessen, S.C. and Brisbois, R.G. (2001) J. Am. Chem. Soc., 123, 3818–3819. (f) Drain, C.M. and Lehn, J.-M. (1994) J. Chem. Soc. Chem.

Coordination Cages

Commun., 2313–2315. (g) Slone, R.V. and Hupp, J.T. (1997) Inorg. Chem., 36, 5422–5423. (h) Stang, P.J., Fan, J. and Olenyuk, B. (1997) Chem. Commun., 1453–1454. (i) Würthner, F. and Sautter, A. (2000) Chem. Commun., 445–446. (j) Yamanoi, Y., Sakamoto, Y., Kusukawa, T., Fujita, M., Sakamoto, S. and Yamaguchi, K. (2001) J. Am. Chem. Soc., 123, 980– 981. (k) Aoyagi, M., Biradha, K. and Fujita, M. (1999) J. Am. Chem. Soc., 121, 7457–7458. (l) Aoyagi, M., Tashiro, S., Tominaga, M., Biradha, K. and Fujita, M. (2002) Chem. Commun., 2036–2037. (m) Tominaga, M., Tashiro, S., Aoyagi, M. and Fujita, M. (2002) Chem. Commun., 2038–2039. (n) Yamaguchi, T., Tashiro, S., Tominaga, M., Kawano, M., Ozeki, T. and Fujita, M. (2004) J. Am. Chem. Soc., 126, 10818–10819. 15 (a) Olenyuk, B., Levin, M.D., Whiteford, J.A., Shield, J.E. and Stang, P.J. (1999) J. Am. Chem. Soc., 121, 10434–10435. (b) Olenyuk, B., Whiteford, J.A., Fechtenkötter, A. and Stang, P.J. (1999) Nature, 398, 796–799. (c) Abrahams, B. F., Egan, S.J. and Robson, R. (1999) J. Am. Chem. Soc., 121, 3535–3536. (d) Moulton, B., Lu, J., Mondal, A. and Zaworotko, M.J. (2001) Chem. Commun., 863–864. (e) Eddaoudi, M., Kim, J., Wachter, J.B., Chae, H.K., O'Keeffe, M. and Yaghi, O.M. (2001) J. Am. Chem. Soc., 123, 4368–4369. (f) Ke, Y., Collins, D.J. and Zhou, H.-C. (2005) Inorg. Chem., 44, 4154–4156. (g) Tominaga, M., Suzuki, K., Kawano, M., Kusukawa, T., Ozeki, T., Sakamoto, S., Yamaguchi, K. and Fujita, M. (2004) Angew. Chem. Int. Ed., 43, 5621–5625; Angew. Chem., 116, 5739–5743. (h) Tominaga, M., Suzuki, K., Murase, T. and Fujita, M. (2005) J. Am. Chem. Soc., 127, 11950–11951. (i) Sato, S., Iida, J., Suzuki, K., Kawano, M., Ozeki, T. and Fujita, M. (2006) Science, 313, 1273–1276. (j) Aoki, S., Shiro, M. and Kimura, E. (2002) Chem. Eur. J., 8, 929– 939. (k) Argent, S.P., Adams, H., RiisJohannessen, T., Jeffery, J.C., Harding, L.P.

References

16

17

18

19

20 21

22

23 24

25 26 27

28

and Ward, M.D. (2006) J. Am. Chem. Soc., 128, 72–73. Kusukawa, T. and Fujita, M. (1998) Angew. Chem. Int. Ed., 37, 3142–3144; Angew. Chem., 110, 3327–3329. Kusukawa, T., Yoshizawa, M. and Fujita, M. (2001) Angew. Chem. Int. Ed., 40, 1879–1884; Angew. Chem., 113, 1931–1936. Gokel, G.W. (1991) Crown Ethers and Cryptands, Royal Society of Chemistry, Cambridge. Szejtli, J. (1988) Cyclodextrin Technology, Kluwer Academic Publishers, Dordrecht. Gutsche, C.D. (1989) Calixarenes, Royal Society of Chemistry, Cambridge. Asfari, Z. Bohmer, V. Harrowﬁeld, J. Vicens, J. and Saadiouli, M. (eds) (2001) Calixarenes, Kluwer Academic Publishers, Dordrecht. Lagona, J., Mukhopadhyay, P., Chakrabarti, S. and Isaacs, L. (2005) Angew. Chem. Int. Ed., 44, 4844–4870; Angew. Chem., 117, 4922–4949. Kusukawa, T. and Fujita, M. (2002) J. Am. Chem. Soc., 124, 13576–13582. Ibukuro, F., Kusukawa, T. and Fujita, M. (1998) J. Am. Chem. Soc., 120, 8561–8562. Yoshizawa, M., Tamura, M. and Fujita, M. (2004) J. Am. Chem. Soc., 126, 6846–6847. Kusukawa, T. and Fujita, M. (1999) J. Am. Chem. Soc., 121, 1397–1398. Yoshizawa, M., Kusukawa, T., Kawano, M., Ohhara, T., Tanaka, I., Kurihara, K., Niimura, N. and Fujita, M. (2005) J. Am. Chem. Soc., 127, 2798–2799. Tashiro, S., Tominaga, M., Kawano, M., Therrien, B., Ozeki, T. and and Fujita, M. (2005) J. Am. Chem. Soc., 127, 4546–4547.

j313

29 Sun, W.-Y., Kusukawa, T. and Fujita, M. (2002) J. Am. Chem. Soc., 124, 11570–11571. 30 Nakabayashi, K., Kawano, M., Yoshizawa, M., Ohkoshi, S. and Fujita, M. (2004) J. Am. Chem. Soc., 126, 16694–16695. 31 Nakabayashi, K., Kawano, M., Kato, T., Furukawa, K., Ohkoshi, S., Hozumi, T. and Fujita, M. (2007) Chem. Asian J., 2, 164–170. 32 (a) Yoshizawa, M., Takeyama, Y., Kusukawa, T. and Fujita, M. (2002) Angew. Chem. Int. Ed., 41, 1347–1349; Angew. Chem., 114, 1403–1405. (b) Takaoka, K., Kawano, M., Ozeki, T. and Fujita, M. (2006) Chem. Commun., 1625–1627. 33 Yoshizawa, M., Takeyama, Y., Okano, T. and Fujita, M. (2003) J. Am. Chem. Soc., 125, 3243–3247. 34 Nishioka, Y., Yamaguchi, T., Yoshizawa, M. and Fujita, M. (2007) J. Am. Chem. Soc., 129, 2000–2001. . 35 Kusukawa, T., Nakai, T., Okano, T. and Fujita, M. (2003) Chem. Lett., 32, 284–285. 36 Yoshizawa, M., Tamura, M. and Fujita, M. (2006) Science, 312, 251–254. 37 Yoshizawa, M., Miyagi, S., Kawano, M., Ishiguro, K. and Fujita, M. (2004) J. Am. Chem. Soc., 126, 9172–9173. 38 Ito, H., Kusukawa, T. and Fujita, M. (2000) Chem. Lett., 29, 598–599. 39 Yoshizawa, M., Sato, N. and Fujita, M. (2005) Chem. Lett., 34, 1392–1393. 40 Yoshizawa, M., Kusukawa, T., Fujita, M. and Yamaguchi, K. (2000) J. Am. Chem. Soc., 122, 6311–6312. 41 Yoshizawa, M., Kusukawa, T., Fujita, M., Sakamoto, S. and Yamaguchi, K. (2001) J. Am. Chem. Soc., 123, 10454–10459.

j315

9 Anion-binding Macrocycles Evgeny A. Katayev, Patricia J. Melfi, and Jonathan L. Sessler

9.1 Introduction

The synthesis and subsequent characterization of macrocyclic compounds is often accompanied by the observation of small species, such as cations, anions, or neutral molecules, encapsulated within the inner cavity of the macrocycle. Anions are particularly important in the context of such recognition because of the range of applications where macrocycles can be used as anion-binding systems, namely in analytical chemistry, biology, medicine, or environmental remediation [1]. The last decade has seen an explosion of interest in the area of anion binding, and it is becoming a common procedure to test new macrocyclic systems for the ability to bind anionic species [2]. Moreover, the range of anions studied has grown as the number of anion-binding agents has increased and their potential utility has come to be better appreciated [3–5]. Not surprisingly, the structures of the macrocyclic receptors in question have grown in their complexity; many of these are so-called hybrid systems that combine various anion-binding units, each of which either provides a speciﬁc type of interaction with the anion (i.e. hydrogen bonding, electrostatic interactions, p-bonding, etc.) or plays a speciﬁc role in the recognition process [6]. This chapter is devoted to recent approaches to the design of such hybrid macrocyclic anion receptors. The ﬁrst stage in the design of synthetic anion-binding receptors involves identifying and characterizing different potential binding motifs. The second stage then involves combining one or more of these motifs in such a way that the resulting receptor is appropriately tailored to the geometry of a given target anion. This tailoring, which in favorable cases provides good substrate-receptor compatibility, allows for the production of macrocyclic systems that, as a general rule, display enhanced afﬁnity for anions relative to open-chain analogs or cyclic compounds that have not been appropriately optimized. The result, at least in principle, is receptors with very high afﬁnities and selectivities for the targeted anionic analytes [2]. However, this idealized goal of substrate-receptor complementarity is rarely achieved,

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

316

j 9 Anion-binding Macrocycles especially in the case of non-spherical anions. For this reason, there is an ongoing search for both new binding motifs and improved design strategies that will allow these motifs to be exploited in promising ways. Over the past few years, a number of reviews has appeared that highlight the development of anion receptors based on widely used anion-binding motifs, including polyammoniums, which are generally considered to be the ﬁrst well-deﬁned anion receptor motifs [7–10], amides [11,12], guanidiniums [13,14], imidazoliums [15], steroids [16], pyrroles [17,18], and also metal-based systems [19,20]. While great success has been encountered using these classic single-motif approaches, a new generation of macrocyclic anion receptors has recently emerged in the literature, wherein a number of different motifs are combined together. These integrated macrocycles represent a novel class of hybrid organic compounds, in which the combination of different motifs is designed to stabilize a speciﬁc interaction between the host and the guest. On some level, therefore, the move toward hybrid anion receptors reﬂects an appreciation of Nature, wherein complex protein-based systems exploit multiple and generally distinct subunits and interactions to achieve exquisite anion-binding selectivity. Scheme 9.1 summarizes the main directions in the development of hybrid macrocycles that are considered most promising in the context of the paired goals of achieving improved anion-binding efﬁciency and enhanced selectivity. These can be viewed in the context of three dominant themes: (a) functionalization of scaffolds, namely the introduction of additional binding motifs with the intent to ﬁne-tune the binding properties of a particular receptor type or to introduce a signaling group for sensing anions [21,22]. (Note: In the context of this chapter, scaffolds will be deﬁned as receptors with similar binding motifs.) (b) receptors for ion pairs – ditopic receptors capable of binding a cation and an anion simultaneously [1,2], and (c) the development of macrocyclic ligands containing different binding motifs for the speciﬁc binding of non-spherical anions [23]. Here, as a general rule, the geometry of the receptor and nature of its presumed non-isotropic interactions with the targeted guest play a crucial role in selective recognition. It is worth mentioning that these three directions overlap signiﬁcantly. They also have connections to other broader themes in supramolecular chemistry. For example, anion-templated assembly, which touches on the emerging theme of dynamic combinatorial chemistry, Functionalization of scaffolds Receptors for ion pairs Ditopic receptors

Receptors with different binding sites

Anion template-based assembly Scheme 9.1 Scheme showing the interplay between different synthetic approaches and the resulting elaborated receptors discussed in this chapter. For the sake of coherence and brevity, the focus is on so-called hybrid macrocyclic anion receptors as detailed further in the text.

9.2 Ditopic Receptors and Receptors for Ion Pairs

represents an effective, albeit still infrequently employed, means of producing highly selective receptors. Given this, it is considered probable that the investigation of assemblies as well as template-based synthesis of known receptors can provide clues to the synthesis of an ideal receptor [2,24,25]. For instance, by using a template it is possible, at least in principle, to direct combinatorial libraries toward the production of highly selective receptors. Such strategies can also provide key insights into the binding properties of a given putative anion recognition motif. On the other hand, to be most informative, studies of this type should involve systems with more than one anion-binding motif, referred to as hybrid receptor systems in the context of this chapter. This chapter consists of two main subchapters in accordance with Scheme 9.1. This scheme is meant to show in a simpliﬁed fashion how the generalized set of macrocyclic receptor systems containing more than one kind of binding motif, socalled hybrid receptors, can be organized according to desired function, namely recognition of more than one substrate (ion pair and ditopic receptors) or the binding of (primarily) a single set of guests (e.g., anions) using multiple recognition motifs (receptors with different binding sites). Needless to say, the dynamic nature of the generalized ion recognition ﬁeld means there is some overlap between these two limiting categorizations. In fact, the same strategies (e.g., functionalization of known scaffolds and template-based assembly) are often used to obtain both classes of receptor, a point that is underscored in Scheme 9.1. Nonetheless, for the sake of coherence, this chapter is organized according to receptor type following the distinctions made by the authors of the original research publications being reviewed. Needless to say, the authors apologize in advance for any confusion this choice of organization might cause, as well as for any omissions to speciﬁc work as necessitated by the length considerations of this chapter.

9.2 Ditopic Receptors and Receptors for Ion Pairs

Since the properties of simple macrocyclic ligands, such as crown ethers, calixarenes, or cholapods, are well known, it is perhaps not surprising that many researchers have chosen to functionalize them with different anion-binding motifs with a view to generating ditopic receptors and receptors for ion pairs. In general, three basic strategies have been employed to effect such functionalization. The ﬁrst involves appending one or more functional groups to the scaffold either to increase the afﬁnity or to enhance the utility of the recognition process. Thus, in the broadest sense this approach includes the tethering of new binding motifs and the attachment of signaling groups that permit sensing applications. The second strategy involves the capping of the basic receptor core with an additional linker (strap) so as to control the conformation, enforce a speciﬁc geometry, or provide additional binding motifs. Since the length of the straps can be adjusted, at least in principle, this strategy often permits a high degree of design optimization. The third strategy consists of linking two or more receptor systems, either open-chain or macrocyclic,

j317

318

j 9 Anion-binding Macrocycles via appropriately chosen linkers so as to generate elaborated polytopic binding agents. Since the individual components in these latter systems need not be the same, considerable selectivity in tuning is possible. For this reason, this third strategy, along with the second, has often been exploited to generate ion pair receptors in addition to systems targeting just anionic (or cationic) substrates. 9.2.1 Crown Complexes

Bradley Smith and coworkers have developed a strapped diazacrown complex, 1 (Scheme 9.2), capable of extracting salts as contact ion pairs [26–30]. In initial work with this receptor, they demonstrated transport of KCl across a phospholipid bilayer membrane [27]. They then demonstrated the transport of an array of ion-pairs across a supported liquid membrane (SLM) [28]. This group noted that transport was more efﬁcient using their ditopic system than with a corresponding set of paired anion and cation receptors under the same conditions. Under the interfacial conditions employed in these experiments, the anion transport selectivity was found to be I > Br > Cl with Kþ. However, when the conditions were altered, (i.e., from through-membrane transport to solid-liquid extraction), the selectivity was reversed. Speciﬁcally, under a solid-liquid transport scenario, extraction with lithium was reported to be 15 times more effective than with either Kþ or Naþ when Cl was used as the counter ion and 11 times better when Br was employed as the counter ion [29]. The authors rationalized this reversal in terms of the higher solubility of the lithium salts, among other reasons. While initial characterization studies of this receptor were focused on metal cation extraction, the anions employed being limited to the halides, more recently Smith and coworkers have investigated oxoanions to good effect. In the context of these latter efforts, a series of crystal structures was compared. In a ﬁrst set of comparisons, employing the KNO3, NaNO3, and LiNO3 salts, an immediate correlation between the size of the cation and the inclusion of the anion within the macrobicycle was seen. The orientation of the NO3 was also

O O

O NH

O

O

O

O

N

HN

NH

HN

O

N H HN N

N H

O

N O

N

R=

O

O

R

O

O O

1

NH

O N H

NH

O

N

H N

Scheme 9.2 Crown-based ditopic receptors.

N

4

O NH

NH

3

2

O

O

Fe O

N

N H H N

HN

N

HN

O

O

O

O

5

N

9.2 Ditopic Receptors and Receptors for Ion Pairs

observed to change, presumably allowing for better alignment of the amide NH protons with the lone pairs of an oxygen atom. Additionally, in all three instances, the amide NH protons were found to form hydrogen bonds with just one of the NO3 oxygen atoms; again, this could be rationalized in terms of the stronger hydrogen bond formed upon ideal alignment of the amide NH protons with the oxygen lone pairs. Similar asymmetry and hydrogen-bond formation were found in a complex formed from NaNO2. In all but one instance (KNO3), some degree of chelation with the counter cation was observed. NMR spectroscopic studies provided evidence for anistropic shielding, a feature that was attributed to the oxoanions. While the NH signals moved downﬁeld upon the addition of NaCl, NaOAc, and KOAc, as would be expected for normal hydrogen bond-mediated anion recognition, this did not prove to be the case for NO3 or NO2. Here, various effects, including an upﬁeld shift in the pyrrolic NH proton signal, were seen in the case of the nitrate and nitrite salts. The authors attributed this behavior to the dual shielding and deshielding properties of NO3 and NO2 [30]. Gunning et al. have investigated the ability of a cyclene-dibenzacrown hybrid macrocycle to bind various substituted phosphates [31]. This system, which functions as an ion pair receptor, consists of a cyclene and dibenzo-24-crown-8 macrocycle joined by a glycine bridge. The zinc complex (Zn-22þ) was readily synthesized and subsequently reacted with either sodium or potassium phosphate salts. The ability to interact with inorganic phosphate and other substituted phosphates was investigated by 1H NMR spectroscopy, UV-vis spectroscopy, and isothermal titration calorimetry (ITC). A Zn2þ-cyclene macrocycle was used in these studies as a control system. ITC experiments revealed 2:1 binding values with dihydrogen phosphate on the order of 4 · 104 M2 and 9 · 104 M2 (in the case where the counter cations were Naþ and Kþ, respectively). The stronger afﬁnities observed for Kþ over Naþ for receptor Zn22þ led the authors to suggest that Kþ is bound by the crown and that the system is functioning as an ion pair receptor. In contrast, the use of Kþ or Naþ salts was observed to have little effect compared to the control zinc complex, which displayed 1:1 binding constants of 1:6 · 104 M1 and 1.5 · 104 M1 for Naþ and Kþ, respectively. Association constants for the substituted phosphates, glycerophosphate monoanion, and 4-nitrophenylphosphate dianion were reported, but were found to be similar to those seen in the case of the control Zn2þ-cyclene system. The thermodynamic quantities DG, DH, and TDS were also reported, and the authors noted that anion-binding with the Zn2þ-cyclene system was exothermic, whereas with the hybrid macrocycle, anion-binding was endothermic. The authors rationalized this dichotomy in terms of the suggestion that a pocket of water exists and that these solvent molecules are hydrogen-bonded to the free receptor system. Upon anion association, this pocket of water must be removed, requiring energy. Nonetheless, since the entropy of the overall system increases, increases in afﬁnity are observed. This work was later expanded to include a larger crown ether system functionalized with either the above-mentioned cyclene macrocycle (2), two cyclene macrocycles (4), or a triazacyclononane macrocycle (3) [32]. Unfortunately, the Cu32þ complex did not display any appreciable afﬁnity for the phosphate species (NaH2PO4, KH2PO4, sodium glycerophosphate, or disodium 4-nitrophenylphosphate).

j319

320

j 9 Anion-binding Macrocycles The authors rationalized this ﬁnding on steric grounds, noting that the phosphate anion was a poor geometric complement to the presumed distorted ﬁve-coordinate copper complex. The Zn42þ complex was shown by UV-vis spectroscopic titrations to coordinate two phosphate anions. Unfortunately, a binding constant could not be determined by ITC. However, comparisons between the free receptors 2 and 4 could be made in the case of the substituted phosphate salts, sodium glycerophosphate and disodium 4-nitrophenyl phosphate. Both receptors formed 1:1 complexes with these anions; however, Zn42þ displayed association constants that were roughly three times greater than those seen in the case of Zn22þ. Intrigued by these results, Gunning investigated binding cooperativity in a similar system in which the cyclene was appended to a 19-benzocrown-6 [33]. Following a procedure ﬁrst put forth by Jencks [34], positive or negative cooperativity of the host-guest system, deﬁned by DGs, was determined for the Naþ, Kþ, and Liþ dihydrogen phosphate salts. This was done by subtracting the DG observed when both the anion (A) and cation (C) were allowed to coordinate concurrently (DGAC) from those DG determined for exclusive anion (DGA) or cation binding (DGC); in other words, DGs ¼ DGA þ DGC DGAC, which is a deﬁnition that permits both positive and negative cooperativity (i.e., DGs may be positive or negative). While the perchlorate salts could be used directly for the analysis of cation binding, the low solubility of the tetrabutylammonium (TBA) dihydrogen phosphate complex required the use of a model system to obtain the baseline dihydrogen phosphate binding values. However, subject to such caveats, DGs values of þ0.83, þ0.72, and þ1.03 (all in kcal mol1) could be calculated for NaH2PO4, KH2PO4, and LiH2PO4, respectively. This system was thus claimed as one of the ﬁrst examples in which positive ion pair cooperative binding was observed. In a related series of investigations, Suksai, Tuntulani, and coworkers appended an amidoferrocene moiety to a diaza-18-crown-6 and probed the anion association constants in the presence of various alkali metal ions [35]. In an effort to assess the utility of this receptor as a potential heteroditopic receptor, they ﬁrst determined the association constants for binding Naþ and Kþ (studied as the perchlorate or hexaﬂuorophosphate salts, respectively); this was done via 1 H NMR spectroscopic titrations carried out in 5% CD3CN/CDCl3. Separate studies of anion binding were carried out using the TBA salts of Cl, Br, and I in the same solvent. However, a Ka value could only be determined in the case of Cl (those for the Br and I salts proved to be too small). Based on these analyses and a comparison of the diaza-18crown-6 as a reference receptor and the functionalized amidoferrocene receptor 5, it was concluded that the chloride afﬁnity decreased when a crown ether was appended onto the core ferrocene receptor. This behavior was rationalized in terms of the enhanced rigidity of the new receptor, which was considered likely to inhibit binding. Addition of TBA chloride to either the Kþ or Naþ complex (previously formed through addition of either KPF6 or NaClO4) of the receptor produced no discernible change in the amide NH signals in the 1 H NMR spectrum up to one equivalent. In the presence of more than one equivalent of Cl, a signiﬁcant shift in the NH signal was observed, which the authors attributed to formation of an ion pair that does not involve binding of the amide NH protons. Similar behavior was observed upon the

9.2 Ditopic Receptors and Receptors for Ion Pairs

addition of TBABr to the Kþ complex. However, upon formation of the presumed Na-Br, Na-I, or K-I pairs, a signiﬁcant downﬁeld shift was seen for the amide NH proton signal in the 1 H NMR spectrum. Furthermore, in these cases, the anionbinding constants (which were previously too low to determine) were found to be greatly enhanced. For instance, the Naþ complex displayed a binding constant of 16 000 M1 for bromide anion, while the K-I and Na-I pairs displayed anion afﬁnities of 39 and 93 M1, respectively. While the latter values are not extremely high, the fact that they are measurable is fully consistent with the proposal that positive cooperativity effects are contributing to ion pair binding in this heteroditopic system. 9.2.2 Calixarenes

Calixarenes and their derivatives (Scheme 9.3) are among the most widely used scaffolds in anion-binding chemistry. Strategies for their functionalization have been reviewed recently [36,37]. The popularity of calixarenes reﬂects to a large extent the ease with which the cores themselves can be modiﬁed as well as the fact that they R2 O R3 O

R3 O

R3 O

3

R OH OH O

R3 R3 O O

O

O

R1

R1

R1

R1

O

R2 R1

R1

R1

R2 R2

O

R1

R2 R2

R2

O

7

6

O R2 8

NO2

NO2

H2N NH

HN

NH

HN

O

NH O

∗

S

O

O

NH

HN

NH O

R

R

O

O

∗

O

O

OH OH O

H2C R2 = (CH2)3(CCH2CH2)3OCH3

NH

OH OH O

9

NH2

R1 =

HN

HN

HN

O

10

Scheme 9.3 Selected representatives of functionalized calixarenes.

j321

322

j 9 Anion-binding Macrocycles possess well-deﬁned upper and lower faces. The most common strategy for calixarene functionalization involves appending additional binding motifs onto the macrocycle to give a product with general structure 6 or 7. Such functionalizations are usually performed so as to provide either two or four groups on the calixarene skeleton. The resulting hybrid systems display binding selectivities that are quite varied. For instance, certain functionalized calixarenes bearing amido substituents on the upper rim displayed a preference for hydrogen sulfate anions (Ka ¼ ca. 105 M1, in CDCl3) [38], whereas others showed selectivity for carboxylate anions (Ka values up to 106 M1) [39–43]. Examples of functionalized calixarenes bearing a charged scaffold, obtained by the use of amidinium [44,45] or imidazolium functionalizing groups [46], are also known in the literature [36] and have been seen to display afﬁnities Ka up to 104–105 M1 in highly solvated solvents, such as DMSO, methanol, and water. A tetraamidinium functionalized, bowl-type cavitand (receptor 8) was developed by Diederich and Sebo [47]. This receptor was found to complex 1,3-dicarboxylate anions with good selectivity and a 1:2 binding stoichiometry both in CD3OD and D2O, as revealed by standard Job plot analysis. In contrast, various nucleotide phosphates were found to be bound with a 1:1 stoichiometry in D2O. In the case of the adenosine phosphates, the association constants increased as a function of nucleotide charge [i.e., the afﬁnity order (Ka, M1) was: cAMP (1400) < AMP (10 000) < ADP (48 700) < ATP (660 000) in pure D2O]. This charged receptor (8) also showed moderate selectivity towards AMP (Ka ¼ 10 000 M1) over other nucleotide monophosphate anions, such as GMP (Ka ¼ 5200 M1), CMP (Ka ¼ 3500 M1), TMP (Ka ¼ 5900 M1), and UMP (Ka ¼ 3800 M1) in D2O containing TRIS buffer (2.5 mM, pH 8.3). The doubly functionalized calix[4]arene 9, bearing two chiral arms, was prepared recently [48]. This neutral system proved not only to be a good receptor for a-phenylglycine anions, it was also found to display good enantioselectivity towards the L-isomer Ka(L)/Ka(D) ¼ 4.76. It was thus proposed by the authors, that this presumably preorganized receptor could function as a chromogenic sensor for nonracemic a-phenylglycine anions. Another strategy for functionalizing calixarenes is to strap the macrocycles. Several systems bearing a diamido strap have been prepared and have shown a selectivity for acetate anions (e.g., receptor 10; Ka ¼ 103 M1 in CDCl3) [49]. Similar selectivity preferences were observed for thiacalix[4]arene receptors functionalized with four ureido or thioureido units on the lower rim [50]. Appropriately strapped calixarenes have also been shown to complex anions and cations simultaneously. Such a system is the ditopic, calixarene-based receptor 11 (Scheme 9.4) reported by Tumcharern and coworkers [51]. Receptor 11, in which the strap incorporates two different binding motifs (urea and amide), showed selectivity for the tetrahedral phenylphosphinate anion (PPhHO2) over simple Y-shaped anions, such as acetate. After initial screening studies, quantitative analyses were carried out using 1 H NMR spectroscopic methods. Speciﬁcally, titrations were carried out in CD3CN using the TBA salts of three promising anions, namely acetate, phenylphosphinate, and diphenylphosphate. Standard analyses of the

O

O

NH

11

O

HN

O

O

O

HN

HN

S

Ph O

Cl

O

N

Ph

NH

12

Cl

N

O

HN

N

Scheme 9.4 Calixarene-based ditopic receptors.

O

O

NH

NH

S

HN O

Cl

O N

O Ph MeO2C O

O

N

N

O

O

HN

HN

13

O O

O

O

NH

NH

N

N

O

O

CO2Me

O

NH

O

O

O

O

O O

14

O

O

HN

O

NH

HN

NHHN

O O

O

R O

15 R = H 16 R = (CH2CH2O)3CH2CH2

O

R O

9.2 Ditopic Receptors and Receptors for Ion Pairs

j323

324

j 9 Anion-binding Macrocycles anion-dependent changes in the thiourea NH proton signals led to calculated 1:1 binding constants of 11 000, 24 000, and 1800 M1, respectively, for these three anions. Separate titration-based studies were carried out to determine the binding constants for the complexation of the alkali metal cations Liþ, Naþ, and Kþ (studied as perchlorate salts). The resulting values led to the suggestion of a clear preference for the sodium cation, with Ka values of 760, 5500, and 970 M1 determined for Liþ, Naþ, and Kþ, respectively. In a third experiment, acetate or diphenylphosphate anions (as their TBA salts) were added to the sodium salt of the receptor. No change in the 1 H NMR spectral features was observed upon the addition of acetate anions until 1 equivalent had been added. Based on this observation, the authors suggested that the Naþ cation was sequestered out of the receptor, forming what was presumably a more stable ion-pair in solution. On the other hand, with diphenylphosphate, 1:1 binding is observed in the presence of Naþ, with the resulting Ka value (for anion complexation) being on the order of 2200 M1, a ﬁnding that could be interpreted in terms of effective ion pair recognition. Thus, the receptor 11, in the absence of Naþ cation, shows selectivity for acetate anions over diphenylphosphates, but in the presence of Naþ this selectivity is reversed. Tumcharern and coworkers postulated that the phenoxy groups may interact favorably with the bound cation and that this could lead to the observed strong binding of the ion pair [51]. Cryptand 12 was synthesized through an amidation reaction in which a calix[3] arene was capped with a tren residue [52]. The anion-binding properties of this receptor, termed N7-azacalix[3]cryptand, were studied in DMSO using UV-vis spectroscopy and the TBA salts of the test anions in question. On the basis of these studies, it was found that halide anions are generally bound more effectively than Y-shaped or trigonal planar anions; in fact, chloride displayed the highest constant for this system with a log K of 4.55 being derived from quantitative analyses. The other anions were bound in the following sequence: Cl > Br > CH3COO > F I > PhCOO > NO3, with the value for perchlorate being too low to be determined. Interestingly, when this system was investigated as a potential ditopic receptor through the addition of Zn2þ (as the triﬂuoromethanesulfonate salt), the afﬁnity for Cl was found to be reduced, while the association constants corresponding to the binding of the other halide anions were found to be enhanced. The authors reasoned that electrostatic interactions involving the bound anion and the zinc cation, while beneﬁcial in the case F, I, and Br, are actually destabilizing in the case of chloride anion since the number of hydrogen bonds between the bound chloride anion and the receptor would be reduced. While there is currently no independent veriﬁcation for this supposition, it nonetheless serves to highlight the fact that there are often several competing effects that serve to modulate the overall binding and selectivity process. Recently, the effect of two proximate cations on anion-binding was investigated [53]. Toward this end, the basic calix[4]arene framework was functionalized on the lower rim with two methyl ester units and in two separate positions with two urealinked bipyridine-functionalized polyether units to give 13. This receptor was designed to allow for the coordination of hard metal cations via the ester groups, as

9.2 Ditopic Receptors and Receptors for Ion Pairs

well as soft cations through the bipyridine units. It was further postulated that various targeted anions would bind to the urea moieties, while being encapsulated between the two co-bound metal cations. Anion afﬁnities for receptor 13 were calculated from 1 H NMR spectroscopic titrations carried out in CDCl3/CD3CN, and were found to be very low for the anion alone. Speciﬁcally, log Ka values were found to be 1.88 for NO3 and 1.4 for CF3SO3, (studied as the corresponding TBA salts). However, in the presence of either sodium or silver cations (added as the tetrakis[3,5-bis(triﬂuoromethyl)phenyl] borate (TFPB) salts), the association constants increased dramatically. In the case of nitrate, the logKa recorded for the Agþ s13 complex was 3.31, whereas that for the corresponding Naþ s13 derivative was found to be 3.82. Finally, according to design expectations, the anion afﬁnity constants for the binuclear metal complex, Naþ sAgþ s13, were found to be considerably higher than those of the monocationic complexes; speciﬁcally, logKa values of 5.07, 4.7, and 4.28 were derived for NO3, CF3SO3, and BF4, respectively. The hybrid calix[4]arene 14, in which calix[4]diquinone is functionalized with an isophthalamide strap [54], has been reported to act as a contact ion pair receptor. In this case, the neutral receptor does not display a discernible afﬁnity for tetrabutylammonium chloride, as determined by 1 H NMR spectroscopic analyses carried out in CD3CN. Similarly, only a moderate afﬁnity for Kþ (as the PF6 salt) was reported, with no evidence of complexation being observed in the case of Naþ or NH4þ (as the Cl salts), as judged by UV-vis spectroscopic titrations. However, the calculated association constants were found to be substantially enhanced when a cation was added to the receptor. In particular, the association constants (log Ka) for chloride anion-binding were found to be 3.79 in the presence of excess Kþ and above 4 in the case of Liþ, Naþ, and NH4þ. According to the authors, such ﬁndings represent a rare example of cooperative binding being manifest in a system wherein essentially no afﬁnity is seen for the individual free ions. Tongraung, Chantarasiri, and Tuntulani have reported two calix[4]arenes 15 and 16, one with urea appendages and a second with a crown strap in addition to the urea arms. These systems were found to bind tetrabutylammonium dihydrogen phosphate with a Ka 200 M1 in DMSO-d6. In line with the authors expectations, the anion-binding afﬁnities were then found to be enhanced upon the addition of sodium cation. For instance, in the presence of this cation, the association constant for calix[4]arene 16 increased to Ka ¼ 1030 M1 (as the PF6 salt; DMSO-d6 solution) [55]. 9.2.3 Cholapods

Cholapods (Scheme 9.5) are receptors based on a steroid core that feature both electron-withdrawing substituents (to enhance anion-binding) and a C20 side chain (to increase solubility in organic solvents). As might be expected, the rigid backbone present in cholapods makes them potentially ideal scaffolds in that they possess a ﬁxed inner cavity as well as conformational rigidity [16,56,57]. However, only a few

j325

326

j 9 Anion-binding Macrocycles C4H9O

OH HN OH

C5H10 O

O

NH

OH

O

O OH OH

C4H9O O

n

O

HN

O

O C5H10

NH

OH O

OH n

OC4H9 OH

O

OC4H9

17 18 n = 0-2 Scheme 9.5 Cholapod-based macrocyclic receptors.

hybrid macrocyclic receptors based on the cholapod paradigm have been reported in the literature, although recently Davis and coworkers reported a steroid-based cryptand 17, which combines two types of functional groups – hydroxyl and amide. This system contains a relatively small cavity incorporated within a rigid framework that serves to preclude intramolecular hydrogen-bonding interactions and thus favors halide anion recognition [58]. In fact, anion-binding studies, performed using standard 1 H NMR spectroscopic titration methods in CDCl3, revealed that receptor 17 is selective for ﬂuoride anions over other halide anions (Ka ¼ 3220, 990, and 250 M1 for F, Cl, and Br, respectively; all studied as the corresponding TBA salts). Although conceptually similar to 17 in terms of its 3-dimensional design, receptor 18, with 1,4-phenylene spacer groups between two cholic acid subunits, was found to bind b-D-glucoside in CDCl3, with Ka ¼ 103 M1 and a 1:1 binding stoichiometry, and to extract this key biomolecule efﬁciently from aqueous solutions into an organic phase [59]. This change in selectivity toward a larger, neutral substrate was rationalized in terms of the expanded system 18 being appropriately larger than its smaller congener 17. 9.2.4 Pyrroles

Calix[4]pyrroles (Scheme 9.6) are another class of scaffold that has been extensively functionalized [18]. Although a range of modiﬁcations of this easy-to-make core have been carried out, most functionalized pyrrole-containing macrocycles can nonetheless be represented by the generalized structure 19. As a rule of thumb, it is possible to effect functionalization on only one group (R1 or R2), on two opposite groups (R1 and R1), on all of the b-pyrrolic positions (R3), or on one of the meso positions (R1 and R2). An example of the latter approach is embodied in the two kinds of so-called deep or extended cavity calix[4]pyrrole prepared to date [60–63]. In both cases, the size of the cavity was enhanced by appending bulky groups, namely substituted aryls [60– 62] or steroid subunits [63], onto the four meso positions. The introduction of these

9.2 Ditopic Receptors and Receptors for Ion Pairs

R3

R3

R3

R1

R2

R1 N H

NH

R

H N 2

R

R1 R3

19

NH

NH

3

CN

21

R3

NO2

O

O

O O

NH

n H N

H N N H

22

O

O

O

O

N H

CN

NC

R

H N

R2

20

R=

N H

R3 HN

R3 R1

R2

H N

H N N H

N H

23

HN

n

n

n

H N

H N N H

N H

24

Scheme 9.6 Calix[4]pyrrole-based receptors.

or other substituents at the meso position of calix[4]pyrroles has been shown to increase the selectivity for various targeted anions [64]. Other systems, not necessarily involving meso-substitution, have allowed for the analytical determination of anions through the use of appended chromophores or ﬂuorophores [65,66]. A different class of substituted calix[4]pyrroles bears nucleic acid base appendages that were designed to effect the cooperative recognition of a complementary mononucleotide substrate. The resulting systems have been studied as key components in ion-selective electrodes [67], as well as stationary phases in chromatographic separations [68]. A relatively new class of pyrrole-containing macrocycles are the N-confused calix[4]pyrroles (e.g., 20 and 21) [69]. Compared to the traditional calix[4]pyrroles, these congeners display totally different binding modes. Such systems bind anions via three normal NH hydrogen bonding interactions, as well as through an unusual CH-anion contact. Additionally, the presence of the so-called inverted pyrrole does not permit the macrocycle to adopt readily the cone-like conformation usually observed in the case of calix[4]pyrrole-anion complexes. Inverted calix [4]pyrroles also allow for facile and selective chemical modiﬁcations of the receptor skeleton. This latter feature has been exploited to prepare the sensor systems 20 and 21. Both of these show selectivity for acetate over chloride and dihydrogen phosphate anions; the Ka values for acetate, chloride, and dihydrogen phosphate are >106, 319, and 810 000 M1, respectively, for 20, and, likewise for the same anion set, 16 600, < 50, and 430 M1 for 21 (all systems studied in DMSO as their

j327

328

j 9 Anion-binding Macrocycles tetrabutylammonium salts). The ratio of binding constants for the acetate anion over chloride is 50-fold for 20 and 28-fold for 21 [70]. These kinds of systems have also been employed in optical sensor arrays that exploit pattern recognition. An interesting addition to the hybrid macrocyclic family is the strapped calix[4] pyrrole. The ﬁrst example of strapped calix[4]pyrroles was system 22 reported by Lee and coworkers in 2002 [71]. More recently, calix[4]pyrroles have been strapped with mresorcinol-derived diether straps of various lengths (23, where n ¼ 1–3). This latter synthetic work was carried out in order to optimize the length of the strap for binding of halide anions [72] and to produce systems that could be compared effectively with simple calix[4]pyrrole (19, R1 ¼ R2 ¼ CH3; R3 ¼ H) as well as the original esterstrapped calix[4]pyrrole 22. From an analysis of the anion-binding properties, an interesting pattern of binding afﬁnities was inferred. For instance, the use of a tight C-4 ether strap (system 23, n ¼ 1) was found to give rise to an extremely high chloridebinding afﬁnity. By contrast, the use of a C-6 ether strap (23, n ¼ 3) produced a system with an enhanced bromide-binding ability, at least relative to the parent macrocycle calix[4]pyrrole (19; R1 ¼ R2 ¼ CH3; R3 ¼ H and R1 ¼ R2 ¼ CH3; R3 ¼ F). In separate work, C3 C4 and C5 amide cis-strapped calix[4]pyrroles (24, n ¼ 1, 2, and 3) were synthesized, and their anion afﬁnities were reported by Lee, Sessler, and coworkers [73]. While increased anion afﬁnity was observed with the TBA salts of Cl, Br and I as compared to the unstrapped parent calix[4]pyrrole 19, no anionto-receptor size selectivity was found. This behavior led the authors to suggest that the anion binds outside the binding pocket. On the other hand, the fact that a dramatic increase in afﬁnity was observed relative to simple calix[4]pyrrole indicates that, in the case of 24, anion binding likely beneﬁts from interactions involving both the amide and pyrrolic NH protons. The expanded porphyrin sapphyrin 25 represents a different class of pyrrolebased macrocyclic anion receptors (Scheme 9.7). Sapphyrin is among the best studied of all expanded porphyrins and was actually the ﬁrst pyrrolic receptor system in which anion recognition was observed [17]. As often happens in the discovery of new types of binding phenomena, the initial ﬁnding was serendipitous. Speciﬁcally, in an effort to obtain diffraction-grade single crystals, sapphyrin was crystallized as what was thought to be the bis-HPF6 salt. However, the ensuing X-ray diffraction analysis and follow-up control studies revealed that the crystalline material contained a ﬂuoride anion bound inside the inner cavity (as well as a PF6 anion not proximate to the sapphyrin core) [74]. Later, it was found that sapphyrin can efﬁciently bind not only ﬂuoride [75], but also the chloride and phosphate anions [76,77]. Although actually antedating the work involving calix[4]pyrroles, it was likewise found that various functional groups could be appended to sapphyrin. The resulting systems were used for the enantioselective recognition of dicarboxylate anions [78], as well as receptors and solution-phase carriers for nucleotides [79,80] and nucleotide analogs [81,82]. In an attempt to generate hybrid macrocyclic carriers for di- and triphosphate species, various oligosapphyrins, e.g., 26 and 27, were prepared. Among the resulting species, oligosapphyrin 27, which remains multiply protonated at neutral pH, was found to be an effective carrier for nucleotide diphosphates [83].

=

R2

O

O

HN

N

N

N

O

OH

OH

N

NH HN

N H

or

N

OH

NH2 or Me

R2

Scheme 9.7 Sapphyrin-based receptors.

=

R1

25

R2

R1

O

N H N H NH N

O

NH NH

O

N

H N

O

N

N

H N

H N

N H

N H

HN

N

N

HN

N

N

N

HN

NH HN

N H

N H

27

O

26

N

HN HN

N

O

NH

HN

HN

O

29 R =

28 R =

R NH R HN N

O

O

N Zn N

R HN

N

Cl

N

N

R HN

9.2 Ditopic Receptors and Receptors for Ion Pairs

j329

330

j 9 Anion-binding Macrocycles Beer and coworkers have reported selectivity for sulfate anions using a zinc tetraphenylporphyrin scaffold [84]. They synthesized two new receptors, 28 and 29, via functionalization of the phenyl groups with either neutral amide moieties or protonated imidazolium linkages. A number of binding studies was performed: 1 H NMR titrations in DMSO-d6, UV-visible titrations in DMSO and a water-DMSO mixture, electrochemical titrations in acetonitrile, and luminescence titrations in acetonitrile or DMSO. As a general rule, the tetra-imidazolium porphyrin 29 was found to be a better anion receptor than the neutral porphyrin 28. Additionally, UV-vis titrations with receptor 29 revealed a strong preference for the TBA salt of hydrogen sulfate (log Ka > 6 M1 in a 95:5 v/v DMSO: water mixture) when compared to the TBA salts of Cl and H2PO4 (log K 4.2 and 4.8, respectively) or Br, I, ATP2 and SO42 (spectral changes were too small to determine association constants). 9.2.5 Miscellaneous

Gong and Gibb have reported a pyridine-containing macrocycle, 30 (R ¼ H), in which anion binding could be improved via an appropriate choice of counter cation (Scheme 9.8) [85]. Initial experiments with the TBA salts of various anions revealed relatively low binding afﬁnities, with the Ka values (1 H NMR in CDCl3) ranging from O O

O

N

O

N

N

O HN

O O

NH O

O O O O

Bu N

S

M

S

O

O

R = H, CH3OH, CH3Ph

N R 2+ R

R

N S

S M

S N

N

S N

=

O

1, 4 - phenylene 2,7- naphthalenediye

M = Cu(II), Ni(II), Zn(II)

R' N R'

R

N

N

S S Zn S S

S S Zn S S

N

N

M

O

O

H

N

R' R'

M

O

N O

R

R

2,7-naphthalenediye (33) 1,1'-biphenyl-3,3'- diyl (34) 1,1'-biphenyl-4,4'-diyl (35)

Scheme 9.8 Ditopic receptors.

R

N H

HN =

N

37

R

R R

32

O

31

30

N

O

O

N Bu

M = Cu(II), Ni(II)

R

N

O

O

S S N Bu

S S M S S

Bu N

NH

38 R = t-Bu, R' = n-C2H13

9.2 Ditopic Receptors and Receptors for Ion Pairs

110 (F) to 32 M1 (I). However, when the counter cations were switched to Lphenylalanine methyl ester (H3NþCH(Bn)CO2Me), these and other anion afﬁnities were found to increase by an order of 103 M1. It should be noted, however, that the inherent anion selectivities remained the same with both counter cations, namely NO3 > Cl > CF3CO2 > Br TsO > I. Ab initio calculations supported the assumption that the binding of both the anion and the cation was cooperative. Enantiomeric recognition was probed in this system via the synthetic addition of a phenyl or alcohol substituent onto the periphery of the macrocycle 30 (R ¼ CH2OH or CH2Ph) [86]. While not the focus of the study, the effects of counter anions (NO3, Cl, Br, TsO) were investigated with the phenyl-functionalized macrocycle. Association constants followed the same trend as the unfunctionalized macrocycle, and unfortunately the inherent enantiomeric selectivity could not be altered by using different anions. Such ﬁndings led the authors to conclude that although this macrocycle is capable of both enantiomeric and ditopic recognition, the two binding phenomena are not necessarily related. Beer and coworkers have published a number of reports [87,88] investigating the use of the dithiocarbamate (dtc) motif in metal-directed self-assembly. While the majority of these reports have either been discussed elsewhere [89] or involved acyclic species and thus fall outside the scope of this Chapter, a few recent publications have been concerned with self-assembled macrocycles [89–91]. In one such paper, the Beer group describe calix[4]arenes functionalized on the upper or lower rims with the dtc ligand, as well as the use of the resulting dimeric structures 31, formed from the addition of Cu2þ, Zn2þ, or Pd2þ, to sense anions electrochemically [91]. Toward this latter end, the Cu(II)/Cu(III) redox couple was monitored as a function of increasing anion concentration, with the degree of cathodic shift being considered a semi-quantitative indicator of anion binding. It was reported that, for equivalent concentrations, the dihydrogen phosphate, benzoate, ﬂuoride, and acetate anions engendered much greater shifts than less charge-dense anions such as hydrogen sulfate or nitrate. Complex 32 was also prepared with differing aromatic spacer groups. While UV-vis titrations in acetonitrile showed little change, a cathodic shift of the Cu(II)/Cu(III) redox couple was observed upon addition of the TBA salt of dihydrogen phosphate or chloride in acetonitrile [92]. In a separate study reported by the Beer group, a similarly functionalized dtc ligand was appended to various aromatic cores, with ﬁve different macrocycles of differing sizes (e.g., 33–36) then being synthesized through a metal-directed selfassembly technique using Zn2þ as the templating cation [90]. The host-guest interactions with 4,4-bipyridine, 1,4-diazabicyclo[2.2.2]octane (DABCO), pyrazine, and the TBA salts of terephthalate and isonicotinate were investigated by 1 H NMR spectroscopic titration methods in a mixture of CDCl3 and DMSO-d6. The authors found that the two anions, terephthalate and isonicotinate, formed strong 1:1 inclusion complexes with all of the receptors produced by this self-assembly process. Molecular modeling calculations provided relative Zn-Zn distances, leading to the conclusion that the terephthalate anion would ﬁt well into receptors 33 and 34. Strong binding was observed with receptors 34 and 36 (>10 000 M1), while receptors 33 and 35 displayed moderate binding constants (5000 and 8000 M1,

j331

332

j 9 Anion-binding Macrocycles respectively). The authors could not rationalize why these larger systems displayed equally strong association constants other than to propose that the ﬂexible nature of the linkers, especially of receptor 36, would allow for variations in the intermolecular Zn-Zn distance. The strongest association constant for isonicotinate was observed for receptor 35 (6500 M1), which contains a biphenyl spacer. Interestingly, control titrations showed no interaction with benzoate, stressing the importance of cooperative interactions with zinc in these systems. Beer and coworkers have also reported a cryptand species formed again from the self-assembly of a dtc ligand with Co(IV) [93]. Addition of the TBA salt of H2PO4 to the resulting receptor in dichloromethane led to a signiﬁcant cathodic shift as compared to what was seen for other anions, such as chloride, benzoate, or hydrogen sulfate [88]. Pyrrole-based dtc macrocycles and cryptands have also been synthesized using copper and nickel cations as the templating species. As a general rule, however, these systems failed to give interpretable results when titrated with H2PO4, Cl, Br, or I as inferred from UV-vis spectroscopic titrations (in CH2Cl2/CH3CN or CH3Cl). However, reliable binding constants could be calculated in the case of a Ni-macrocycle with TBA-OAc (log Ka ¼ 7.06) and a Cu2þ-cryptand with TBA-benzoate (log Ka ¼ 5.73). Still, it is important to note that in the case of the Ni-macrocycle, the logK values did not prove to be substantially higher than those obtained for the control species, nickel diethyldithiocarbamate [94]. Plieger et al. have recently reported a series of salen-based anion receptors that have been modiﬁed in an effort to improve sulfate extraction [95,96]. In earlier work, these researchers had reported that appending tertiary amine groups to a metalated salen scaffold (cf., e.g., receptor 37) led to preferential sulfate binding and extraction over hydrogen sulfate [95]. They then investigated the effect of pre-organizing the anion-binding site by inserting a hexamethylene strap between the two amine groups (to give e.g., receptor 38). While extraction studies led to the conclusion that sulfate binding had increased, it was reported that the sulfate was not encapsulated by the macrocycle; rather, a dimer was formed with two sulfate anions being complexed by two salen ligands. On this basis, it was rationalized that the hexamethylene strap was too small to allow for idealized sulfate anion encapsulation [96].

9.3 Receptors with Different Binding Sites

In this section, we discuss macrocyclic receptors prepared speciﬁcally with the aim of producing molecules with different binding motifs. Such systems, which will be considered mostly in the context of binding anions with multiple protonation sites (e.g., phosphates or sulfates) in solution, can be considered as a subset of the large class of compounds known as ditopic receptors. The combination of amine-containing motifs with amides is a popular approach to the design of hybrid anion-binding agents. The resulting receptors contain two kinds of recognition motifs and, as a result, can act as both hydrogen-bond donors (amide NH protons) and hydrogen-bond acceptors (amine lone pairs). They can also

9.3 Receptors with Different Binding Sites

stabilize a bound anion through a Coulombic interaction involving a protonated (and hence charged) amine group, especially under acidic conditions. Taken together, these characteristics make amido-amine receptors promising for use in both water and organic phases. One of the ﬁrst hybrid amido-amine anion receptors was reported by Lehn in 1988 [97]. This system, the tricyclic receptor 39, possesses three amide and three amine functionalities and was prepared through a coupling of a cyclophane bearing three appended aliphatic amine chains and 1,3,5-benzenetriacyl chloride. By using various aminoalkyl cyclophane derivatives, the size of the resulting cavity could be varied (i.e., n ¼ 1–3 in structure 39), with all three derivatives being shown by NMR spectroscopy to complex nitrate anions efﬁciently, although no quantitative binding data was provided. More recently, a series of amido-amine macrocyclic receptors, 40-43, was developed by Bowman-James et al. (Scheme 9.9). Because of the presence of both binding motifs, these receptors, like many of their congeners, were seen to bind dihydrogen

O

2BPh4-

NH

O

Y

O

NH n(H2C) N

X NH

NH

(CH2)n

(CH2)n

Y

Y Y

Y Y

Y

Y

Y

O O

O

O O

O O

42

Scheme 9.9 Amido-amine type receptors.

O

NH N HN HN NH F N N H N N F NH HN NH N HN

Y

Y = O, X = N Y = S, X = N

HN X

41 Y = O, X = CH Y = O, X = N

HN X

N

NH

X

NH X HN NH HN N N NH

+

N

40 Y = O, X = CH Y = O, X = N Y = S, X = CH Y = S, X = N

39

HN

+

HN

Y

X NH

N

NH

N

Y

HN

N

N

Y

(FHF-)•43

Y

j333

334

j 9 Anion-binding Macrocycles phosphate and hydrogen sulfate anions well in DMSO-d6 with observed Ka values up to 105 M1. These association constants were roughly 2 orders of magnitude greater than those seen for nitrate, chloride, or iodide in the same solvent, and roughly 4 orders of magnitude greater than the Ka values seen for perchlorate [98]. 3-Dimensional analogs of the original amido-amine receptors (e.g., 42) [99,100] were also reported by these same researchers and were put forward as part of an ongoing, rational effort to develop receptors for tetrahedral anions. A single-crystal X-ray diffraction structure of the complex 42 sH2SO4 revealed that sulfate anion binding in the solid state involves eight NH-derived hydrogen-bonding interactions, where two of the NH protons are situated on the amino motifs. In this system, the orientation of the amide protons is presumably pre-established in a favorable orientation as the result of hydrogen-bonding interactions involving the pyridine nitrogen lone pairs. Although increased afﬁnity and selectivity for tetrahedral anionic substrates can be expected for appropriately designed 3-dimensional receptor systems as compared to the corresponding macrocyclic (and hence topographically planar) analogs, in the case of 42 the spherical halide anions, ﬂuoride (log Ka > 5) and chloride (log Ka > 3), compete effectively with the targeted tetrahedral species, namely phosphate (log Ka > 3) and sulfate (log Ka 1.8), at least in DMSO. Bowman-James et al. have also described a new system 43 containing four amidecontaining fragments connected through polyamine linkers. These tricyclic systems possess a unique complementary structure for ditopic anions such as the biﬂuoride anion FHF [101]. In particular, receptor 43 was found to trap efﬁciently a biﬂuoride anion inside the macrocycle and to stabilize a complex with the molecular formula Bu4N[43(FHF)]3H2O. 1 H NMR spectroscopic titration studies of 43 provided support for the proposed selectivity; speciﬁcally, these studies revealed that the FHF anion is bound with a Ka 5500 M1 in DMSO-d6, whereas the dihydrogen phosphate (Ka 740 M1), azide (Ka 340 M1), and acetate anions (Ka 100 M1) are bound far less well. Other anions such as the halides, hydrogen sulfate, nitrate, and perchloride did not show any propensity to bind to this host molecule. An exciting class of amido-amine hybrid anion-binding systems was reported recently by Kubik et al. (Scheme 9.10) [102]. One of the ﬁrst receptors of this type, 45, is derived from L-proline and 6-aminopicolinic acid, and consists of two cyclopeptides bridged by a spacer group. The net result is a potentially ditopic binding system. Earlier studies of the simple cyclopeptide 44 showed that this latter monotopic analog forms a 2:1 complex with sulfate. Thus, it was reasoned that linking two subunits together, to provide systems such as 45-48, would give rise to highly efﬁcient systems that would bind sulfate with a 1:1 stoichiometry. In the case of the prototypical system 45, the stability constants were measured in a 50% methanol/water solution using both 1 H NMR spectroscopic titrations and ITC analysis; the resulting Ka values were found to vary over the 102–105 M1 range and to decrease in the following order (log Ka values in parentheses as determined by NMR/ITC measurements, respectively): SO42 (5.54/4.55) > I (3.95/3.79) > Br (3.72/3.45) > Cl (2.85/2.85) > NO3 (2.11/not determined). A good correlation between the NMR spectroscopic analyses and ITC measurements was observed. The advantage of the latter studies is that they allowed the associated thermodynamic

j335

9.3 Receptors with Different Binding Sites

O

N H

N

N

R=

X

O

O

N

O

O

O

N

NH

O

N H

N H

HN

O

O

O

O NH

HN

HN N

NH N O

O

N

45 X =

S

46 X =

S OH

47 X =

S S

S S

NH NH NH

48 X =

S S

F

NH O

N H

O

F

49

O

H N

F

O

44 X = H R X R

F HN

NH

O

HN

50

HN HN

R N NH

51

O

HN

HN

S S

HN O

N H H N

N NH HN N

N H

N R

H N

52

Scheme 9.10 Selected hybrid receptors.

parameters to be obtained. In particular, for the binding of sodium sulfate, DH was found to be 15.0 kJ mol1 and TDS ¼ 11.0 kJ mol1. In contrast, for sodium iodide, values of DH ¼ 13.2 kJ mol1 and TDS ¼ 8.4 kJ mol1 were measured. A noteworthy ﬁnding emerging from these detailed studies was that the monotopic cyclopeptide 44 actually proved to be a more efﬁcient receptor for sulfate anions than the ditopic system 45 (for 44, DH ¼ 19.3 kJ/mol and logKT ¼ 6.48 in 50% D2O/CH3OH, where KT ¼ Ka1Ka2[M2]). Furthermore, in contrast with the design expectations, microcalorimetric titrations (e.g., ITC analyses) revealed that sulfate is bound to receptor 45 with a 2:1 binding stoichiometry. Needless to say, the ﬁnding that the simple macrocyclic control system 44 was a better receptor than 45 was an unexpected result. The authors attributed this ﬁnding to the low ﬂexibility inherent in the ditopic receptor 45. Hence, by choosing a more appropriate linkage, it should prove possible to enhance the binding afﬁnity. Toward this end, the authors generated a dynamic combinatorial library (DCL); this was done by exposing dimer 44 to several thiols in the presence of oxygen and potassium sulfate using a 2:1 acetonitrile/water mixture as the solvent [103]. Under these conditions, ampliﬁcation of the two bis(cyclopeptides) 47 and 48 was observed, supporting the conclusion that both systems bind sulfate strongly [104]. In agreement with this consideration, receptors 47 and 48 were indeed found to display an afﬁnity for sulfate anion that was enhanced 10-fold relative to what was seen in the case of the ﬁrst generation system 45, at least in a 2:1 acetonitrile/water mixture.

O

336

j 9 Anion-binding Macrocycles Such intriguing ﬁndings provide support for the notion that the right spacer group can signiﬁcantly increase the binding afﬁnity of a given type of ditopic receptor. Another interesting hybrid receptor, system 49, was recently reported by Gale and coworkers [105]. Receptor 49 contains urea and amide motifs in an ideal geometry for binding carboxylates. It was shown by 1 H NMR spectroscopic titrations carried out in DMSO-d6 that this receptor possesses high afﬁnity for acetate (Ka ¼ 16 500 M1) and benzoate (Ka ¼ 6430 M1) anions. It also displays a 100-fold selectivity for these carboxylate anions over dihydrogen phosphate (Ka ¼ 142 M1) or chloride (Ka ¼ 194 M1) anions. Analysis of the 1 H NMR spectroscopic binding data, particularly the chemical shift of each type of proton, led the authors to suggest that the receptor interacts with carboxylate anions differently from the way it interacts with either dihydrogen phosphate or chloride. In the context of the above studies, it was found that receptor 49 was able to ﬁx carbon dioxide from the atmosphere. This unexpected discovery was made during attempts to crystallize the receptor in the presence of tetrabutylammonium ﬂuoride. The resulting single-crystal X-ray diffraction analysis revealed the presence of a carbonate anion coordinated by six hydrogen bonds in the cavity of the macrocycle, the formula of the resulting complex being (49)4[(C4H9)4Nþ]9(CO32)4(F) s4H2O. Shinkai and coworkers have reported that a combination of both allosteric and hydrogen-bonding interactions with anions was achieved by the bicyclic host 50 [106]. In this work, the host molecule, which consists of two bisamide-type subunits separated by a diethynyltetraﬂuorophenyl group, was designed to operate in a turnstile fashion. Speciﬁcally, the central spacer was expected to rotate upon binding of one guest molecule and to elicit the cooperative binding of a second guest molecule. This receptor system was observed to preferentially bind anions such as acetate, (Ka ¼ 144 M1), dibenzyl phosphinate, (Ka ¼ 196 M1) and chloride (Ka ¼ 110 M1) in THF-d8 DMSO-d6, 5:1 v/v. Pyrroles are good hydrogen-bond donors and, subject to appropriate caveats, are relatively insensitive to either protonation or deprotonation effects. Therefore, they have seen increasing use in the construction of highly efﬁcient anion receptors [18], as well as more recently in the construction of hybrid receptors. An example of a pyrrole-containing hybrid receptor that was reported quite recently is compound 51 [107]. This was prepared by condensing 2,5-pyrroledialdehyde with 1,3,5-triethyl2,4,6-trimethylaminobenzene followed immediately by the reduction of the imino bonds to the corresponding amines. This hybrid amino-pyrrole receptor displayed a unique ability to recognize speciﬁcally the b anomer of D-glucose (Ka ¼ 4.83 · 104 M1) and its corresponding alkyl glucosides with complete b/a-selectivity. It was also able to discriminate effectively the b-monosaccharides of the gluco-series from the a and b anomers of the galacto and manno series. A somewhat analogous hybrid macrocycle was developed by Garcia-España and coworkers [108]. Speciﬁcally, these researchers reported a series of [2 þ 2] dipodal systems (52) containing 1H-pyrazole rings and amino linkers. These receptors were found to complex L-glutamic acid efﬁciently in water [109]. The actual structures could be varied somewhat by changing the substituents on the amines (i.e. R ¼ alkyl or benzyl), and the stability of the macrocycle-glutamic acid adduct was found to be

9.3 Receptors with Different Binding Sites

j337

higher when R was benzyl. This result was rationalized in terms of the presence of stabilizing p-cation interactions. Expanded porphyrins can also be considered a class of hybrid receptors. Despite the fact that these ligands were originally explored in the context of metal complexation [110], anion interactions were ﬁrst observed for such systems in 1990, as noted in Section 9.2.4 above [74]. Cyclo[8]pyrrole (53), a recent addition to the expanded porphyrin family, contains both aminic and iminic nitrogen atoms in its core (Scheme 9.11). While analogous in some respects to the sapphyrins discussed earlier (cf. Section 9.2.4), it is discussed here in the context of hybrid receptors because it contains two types of binding motifs within its macrocyclic core, namely imino-like nitrogen atoms that can serve as hydrogen-bond acceptors (or donors after protonation) and pyrrole NH sites that are expected to act as strong hydrogen-bond donors. The combination of these two motifs in one macrocyclic cavity was shown to facilitate the complexation of sulfuric acid [111].

N H

O

HN

NH N H N

NH N

N

HN NH

HN

NH

N

O

N

N

HN

NH

HN N

NH O

53 O

55

O HN N

N NH

HN

NH

HN

O

O NH

N

N

S

HN N

N NH HN

O

O NH

S

N

O

56 Scheme 9.11 Pyrrole-based hybrid receptors.

57

HN N

N NH HN

HN

NH O

O

54

N NH

HN

N

58

338

j 9 Anion-binding Macrocycles In an attempt to tune the anion-binding features of pyrrole-containing hybrid receptors so as to favor tetrahedral species, the authors groups have recently introduced pyridine 2,6-diamido subunits into pyrrolic macrocycles through the use of simple-to-effect iminations [112]. The advantage of this latter approach lies in the combination of known anion recognition groups with an imine functionality that can provide additional anion recognition sites, either through protonation or via direct interaction with the acid form of an oxoanion (e.g., a P-OH proton). The ﬁrst member of this potentially large series of compounds to be reported was receptor 54, a system that was shown to possess a high afﬁnity for both sulfate and dihydrogen phosphate (both studied as the corresponding TBA salts). Whereas the ﬁrst of these anions was bound in a 1:1 fashion (Ka ¼ 64 000 M1 in acetonitrile), dihydrogen phosphate was actually complexed in a stepwise fashion and with a 1:2 stoichiometry (Ka1 ¼ 342 000 and Ka2 ¼ 26 000 M1 in acetonitrile). Moreover, this receptor displayed no appreciable afﬁnity for nitrate, bromide, or perchlorate. This high inherent selectivity makes these systems of potential interest in the area of nuclear waste remediation, where the selective extraction of sulfate from nitrate-rich mixtures is considered beneﬁcial [113]. With the goal of simplifying the binding stoichiometry and enhancing the sulfatephosphate selectivity, a tolyl group was added to the meso-like position of the dipyrromethane fragment. This gave rise to receptor 55, which was designed to be far more rigid than 54 because of a need to avoid potential steric interactions between the tolyl and methyl substituents on the pyrrolic 3-positions [114]. This structural ﬁne-tuning produced a receptor with a selectivity for hydrogen sulfate anion (Ka ¼ 108 000 M1) relative to dihydrogen phosphate (Ka ¼ 29 000 M1) in acetonitrile. In analogy to what proved true for cyclo[8]pyrrole (53) (see above), hybrid system 55 was found to complex sulfuric acid in the solid state through a combination of two imino fragments that become protonated upon complexation, as well as through hydrogen bonds involving the pyrrole and amido motifs. In the context of the above studies, it was discovered that the same kinds of building blocks, namely diformyl pyrrolic precursors and diamines, could be used to effect the combinatorial selection of macrocyclic Schiff base targets when an appropriate anion is used as a template [115]. For example, condensation between a diformyl bipyrrole and a diamine produces the [2 þ 2] macrocycle 56 cleanly, but only when the reaction was carried out in the presence of the acid form of tetrahedral oxoanions. The use of other acids led to formation of a mixture of [n þ n] macrocycles, where n ¼ 1–4, as well as products presumed to be open-chain oligomers, as inferred from mass spectrometric analyses. In acetonitrile, receptor 56 displayed binding properties analogous to those of receptor 54, with the exception that no evidence of chloride binding was seen. A particularly intriguing property of receptor 56 is that it undergoes rearrangement in acetonitrile to produce the corresponding [3 þ 3] macrocycle (instead of the [2 þ 2] congener) in the presence of TBA-HSO4 or TBA-H2PO4 provided that the reaction mixture is allowed to sit for several days in the absence of stirring. Under conditions of greater agitation, the initial [2 þ 2] macrocycle precipitated in the form of its anion complex salt. The molecular structure of the [3 þ 3] product was conﬁrmed via X-ray

9.3 Receptors with Different Binding Sites

diffraction analysis of the sulfuric acid salt; this analysis revealed, not surprisingly, that the bound anion is stabilized in part by hydrogen-bonding interactions involving all three bipyrrole fragments. This group of anion receptors has been expanded by the replacement of the 2,6diamidopyridine with 2,5-diamidothiophene (structures 57 and 58) [116]. Condensation of bis(2-aminophenyl)-thiophene-2,5-dicarboxamide with either a diformyl dipyrromethane or diformyl bipyrrole led to the formation of macrocycles 57 and 58, respectively. Binding constants for the TBA salts of AcO, Cl, Br, HSO4, H2PO4, and NO3 were determined through UV-vis spectroscopic titrations carried out in dichloroethane, and were generally found to be on the order of 103– 104 M1. It was reported that the larger, more ﬂexible receptor 57 displayed overall higher binding afﬁnities when compared to 58, a ﬁnding that was attributed to the ﬂexibility of the receptor, enabling it to easily accommodate anions of different geometries. While still giving rise to Ka values below those of 57, the more rigid receptor 58 did show increased binding afﬁnities for the larger anions, namely bromide and hydrogen sulfate. Binding of a complex anionic species with the help of hybrid macrocycles able to provide different interactions with the anion simultaneously was demonstrated in an elegant manner by Stoddart et al. They reported the organization of macrocycle via the anion-assisted assembly of large [5]- and [6]pseudorotaxanes from dibenzylammonium threads and polyether macrocycles. The resulting systems were found to be interwoven around a central hexaﬂuorophosphate anion (Scheme 9.12, structure 59) [117]. A similar strategy, also permitting the highly efﬁcient synthesis of macrocyclic systems, was recently reported by Beer and coworkers [24]. This work, which involves a general anion templation strategy for the construction of a variety of interpenetrated and interlocked molecular structures, has led to the preparation of a series of [2]rotaxanes and [2]catenanes [118]. The resulting interlocked structures 60-62 feature a unique topology and a deﬁned directionality within the hydrogenbonding network; they also exhibit a high degree of selectivity for chloride. The anion templated syntheses by Beer afforded receptors 60-62 by reacting the chloride salt of the starting pyridinium precursor with an appropriately chosen 1,3diamidobenzene derivative. Based on 1 H NMR spectroscopic titrations carried out in acetone-d6, the structure Cl s60 was found to form with a 1:1 stoichiometry and with an association constant of 2400 M1. It is important to appreciate that these ditopic macrocycles contain a polyether moiety on one side and present a diamido motif on the other. Addition of a charged pyridinium cation to the substrate-free form of these macrocycles then generates a highly efﬁcient receptor for chloride, presumably as the result of both templateinduced organization and electrostatic effects. The further incorporation of hydroquinone groups and polyether functionalities into the cyclic framework presumably contributes to the stabilization of the cationic pyridinium component within the ﬁnal interpenetrated structure. Interestingly, while titration of the pyridinium cation building block alone revealed a preference for oxoanions, the ﬁnal receptor-anion

j339

340

j 9 Anion-binding Macrocycles O

O

H2 N

O Ph

O

O O Ph

O

Ph

O

PF6-

H2N

O O

O

O Ph O

O

Ph

Ph N H2

O O

O

O

O Ph O

O

O

O

O

O N H H ClH H N O N

N

O

H

N H

O

C6H13

O

R

O

O H O O

Cl-•60

O

NH2

Ph

C6H13

O N H H Cl H H N O N R

N

O

H H

O O

O

H

N

O

Cl-•61

59 (PF6-) R=

O

O O

H

H

H

PF6-

PrO PrO

O

PrO

N O N

ON H

ON

H

H

PrO O

ClH

NO

O Cl-•62

OC Cl N OC Re OC N

N

O

H

O N H H O N

H H

H

O O

O

H

N O

O H

H

O

O

O

N H

O

N

O O

PF6-

63

PrO PrO PrO PrO

Scheme 9.12 [2]Rotaxane chloride-selective receptors.

complex PF6 s61 did not display any particular afﬁnity for normal test oxoanions such as dihydrogen phosphate and acetate. However, it was found to bind chloride anions efﬁciently (Ka ¼ 1130 M1 in CDCl3 CD3OD 1:1). More recently, Beer and his group developed a high-yielding anion-templated synthesis of the chloride-selective receptor 62. This was done by carrying out two identical cyclizations of two macrocyclic rings in the presence of chloride anion,

9.4 Conclusions

which served as the templating agent [119]. Here, a key feature was the double ringclosing metathesis (RCM) reaction carried out under conditions where the components of the receptor were orthogonal to each other. This resulted in the formation of a [2]catenane product. Interestingly, the RCM reaction was found to depend critically on the relative concentration of the chloride anion. The resulting receptor was found to bind anions with an apparent 1:2 stoichiometry, a 20-fold selectivity for chloride compared to acetate being observed [K11 and K12 values (all M1 and all in acetoned6) for chloride are 9420 and 160, for acetate 420 and 40, and for bromide 790 and 40 M1]. A new hybrid system 63, selective for hydrogen sulfate, has also been synthesized with the aid of an anion-templating agent [118]. This receptor combines in one framework a vinyl-appended rhenium(I)bipyridyl fragment, a calix[4]arene, four amido fragments, and a pyridinium cation. Luminescent spectroscopy binding studies with chloride, nitrate, and hydrogen sulfate (studied as their hexaﬂuorophosphate salts in acetone-d6) revealed the ability of the [2]rotaxane 63 to operate as a sensor for hydrogen sulfate via a signiﬁcant enhancement in the rhenium bipyridine metal to ligand charge transfer (MLCT) emission intensity. Under these conditions, hydrogen sulfate anions were bound strongly (Ka > 106 M1), with the binding constants for both chloride and nitrate anions also being rather high (e.g., on the order of 105 M1) for these two species, respectively.

9.4 Conclusions

Hybrid macrocycles represent one of the current, cutting-edge approaches to the preparation of functioning anion receptors. While an almost inﬁnite number of hybrid receptor systems is conceivable, within the anion recognition area the focus to date has been on two broad structural classes, namely ditopic receptors, including those for ion pairs, and functionalized systems containing more than one kind of anion-binding motif. The latter systems offer the possibility of bringing more than one key anion-binding motif to bear on a targeted anionic substrate and thus have the potential to display exceptionally high afﬁnities and selectivities. Likewise, systems containing recognition sites for both anions and cations are of interest as receptors for zwitterions and salts. Thus, considered together, these hybrid receptor classes offer the prospect of binding a large range of species and of doing so with considerable selectivity and afﬁnity. At present, this potential is only beginning to be realized. However, the large number of motifs that can be employed, considered in conjunction with the almost inﬁnite number of potentially conceivable structural forms the resulting receptors could attain, means that the opportunities for further development are ample. Indeed, it is safe to predict that the day is not far off when, after the full power of modern theoretical calculations and combinatorial synthetic methods are harnessed, it should be possible to design rationally and synthesize expeditiously a hybrid receptor that is speciﬁcally ﬁne-tuned for the binding and recognition of a speciﬁcally chosen anionic substrate.

j341

342

j 9 Anion-binding Macrocycles 9.5 Experimental: Selected Procedures 9.5.1 Macrocycle H2SO4 s53

A 1-L round bottom ﬂask was charged with a stir bar, dichloromethane (500 mL), and a solution of FeCl3 s6H2O (2.7 g, 10 mmol) in 1-M sulfuric acid (100 mL). The resulting biphasic mixture was stirred at 300 rpm, while a solution of 3,4-dimethylbipyrrole (188 mg; 1 mmol) in dichloromethane (50 mL) was added slowly by syringe pump over a period of 9 h, with the needle injecting into the organic phase. After completion of the addition, the mixture was stirred for an additional 5 h. The phases were then separated, and the organic phase was dried over anhydrous sodium sulfate. After removal of the solids by ﬁltration, the solvent was removed in vacuo to yield the crude product. The crude product was puriﬁed by column chromatography on silica gel using dichloromethane containing methanol (2–5%) as the eluent. The solid residue obtained in this way was recrystallized from dichloromethane/methanol to yield H2SO4 s53 as a dark microcrystalline powder in 74% yield (155 mg; 0.185 mmol). 9.5.2 Macrocycle 55

Bis(2-aminophenyl)pyridine-2,6-dicarboxamide (200 mg, 0.576 mmol), 2,20 -(5-formyl-3-methyl-4-propyl-pyrrolyl)(p-tolyl)methane (232 mg, 0.665 mmol), and 2.2 equiv. of conc. H2SO4 were mixed in dry methanol (50 mL). The mixture was heated to reﬂux under Ar for 15 min. After cooling, the volatile components were removed under low pressure. The resulting solid was dissolved in CHCl3 (10 mL) and layered with pentane (30 mL). The product, macrocycle 55, was found to crystallize overnight in the form of its sulfate salt in 96% yield (454 mg; 0.558 mmol). The corresponding free base form was prepared using the same procedure as that described for H2SO4 s55, with the exception that after heating to reﬂux in methanol for 15 min, an excess (20 equiv.) of triethylamine was added. The resulting mixture was then heated to reﬂux for an additional 5 min and allowed cool to room temperature with stirring for 30 min. The product was collected as a yellow precipitate in 96% yield (399 mg; 0.536 mmol).

Acknowledgments

Gratitude is expressed to the U.S. Department of Energy Ofﬁce of Science (grants Nos. DE-FG02-01ER15186 and DE-FG02-04ER63741 to J.L.S.), the National Institutes of Health (grant No. GM 58907 to J.L.S.), and the Russian Foundation for Basic Research (grant No. 05-03-08017 to E.A.K.) for ﬁnancial support of this work and to Jeong Tae Lee for assistance in the early organizational stages of this project.

References

j343

Abbreviations

AcO ADP AMP ATP cAMP cyclene DMSO DCL dtc ITC MLCT salen RCM TBA tren TFPB TsO

acetate anion (CH3CO2) adenosine diphosphate adenosine monophosphate adenosine triphosphate cyclic adenosine monophosphate 1,4,7,10-tetraazacyclododecane dimethyl sulfoxide dynamic combinatorial library dithiocarbamate isothermal titration calorimetry metal to ligand charge transfer 2:1 Schiff base formed from salicylic aldehyde and ethylene diamine, or their derivatives ring-closing metathesis tetrabutylammonium tris(2-aminoethyl)amine tetrakis[3,5-bis(triﬂuoromethyl)phenyl]borate tosylate anion (p-CH3-C6H4-SO3)

References 1 Bianchi, A., Bowman-James, K. and Garcia-España, E. (1997) Supramolecular Chemistry of Anions, Wiley-VCH, New York. 2 Sessler, J.L., Gale, P.A. and Cho, W.-S. (2006) Anion Receptor Chemistry, in Monographs in Supramolecular Chemistry (ed.Stoddart, J.F.) Royal Society of Chemistry, Cambridge. 3 Gunnlaugsson, T., Glynn, M., Tocci, G.M., Kruger, P.E. and Pfeffer, F.M. (2006) Coord. Chem. Rev., 250, 3094– 3117. 4 Nguyen, B.T. and Anslyn, E.V. (2006) Coord. Chem. Rev., 250, 3118–3127. 5 Martínez-Máñez, R. and Sancenón, F. (2006) Coord. Chem. Rev., 250, 3081– 3093. 6 Gale, P.A. and Quesada, R. (2006) Coord. Chem. Rev., 250, 3219–3244. 7 Llinares, J.M., Powell, D. and BowmanJames, K. (2003) Coord. Chem. Rev., 240, 57–75.

8 Hosseini, M.W. (2003) Coord. Chem. Rev., 240, 157–166. 9 (a) Bowman-James, K. (2005) Acc. Chem. Res., 38, 671–676. (b) Kang, S.O., Begum, R.A. and Bowman-James, W. (2006) Angew. Chem., 118, 8048–8061; Angew. Chem. Int. Ed., 45, 7882–7894. 10 Wichmann, K., Antonioli, B., Söhnel, T., Wenzel, M., Gloe, K., Price, J.R., Lindoy, L.F., Blake, A.J. and Schröder, M. (2006) Coord. Chem. Rev., 250, 2987–3003. 11 Bondy, C.R. and Loeb, S.J. (2003) Coord. Chem. Rev., 240, 77–99. 12 Kang, S.O., Hossain, Md.A. and Bowman-James, K. (2006) Coord. Chem. Rev., 250, 3038–3052. 13 Best, M.D., Tobey, S.L. and Anslyn, E.V. (2003) Coord. Chem. Rev., 240, 3–15. 14 Schmuck, C. (2006) Coord. Chem. Rev., 250, 3053–3067. 15 Yoon, J., Kim, S.K., Singh, J. and Kim, K. S. (2006) Chem. Soc. Rev., 35, 355–360.

344

j 9 Anion-binding Macrocycles 16 Davis, A.P. and Joos, J.-B. (2003) Coord. Chem. Rev., 240, 143–156. 17 Sessler, J.L. and Davis, J.M. (2001) Acc. Chem. Res., 34, 989–997. 18 Sessler, J.L., Camilolo, S. and Gale, P.A. (2003) Coord. Chem. Rev., 240, 17–55. 19 Beer, P.D. and Hays, E.J. (2003) Coord. Chem. Rev., 240, 167–189. 20 (a) Severin, K. (2006) Chem. Commun., 3859–3867. (b) Severin, K. (2006) Synthetic Organometallic Hosts for Biologically Interesting Molecules’’ in Bioorganometallic Chemistry (Topics in Organometallic Chemistry) (ed. Simonneaux, G.), Springer, Heidelberg, 123–142. 21 Martinez-Máñez, R. and Sancenón, F. (2003) Chem. Rev., 103, 4419–4476. 22 Callan, J.F., Prasanna de Silva, A. and Magri, D.C. (2005) Tetrahedron, 61, 8551–8588. 23 Katayev, E.A., Ustynyuk, Y.A. and Sessler, J.L. (2006) Coord. Chem. Rev., 250, 3004–3037. 24 Lankshear, M.D. and Beer, P.D. (2006) Coord. Chem. Rev., 250, 3142–3160. 25 Gimeno, N. and Vilar, R. (2006) Coord. Chem. Rev., 250, 3161–3189. 26 (a) Mahoney, J.M., Beatty, A.M. and Smith, B.D. (2001) J. Am. Chem. Soc., 123, 5847–5848. (b) Mahoney, J.M., Davis, J.P., Beatty, A.M. and Smith, B.D. (2003) J. Org. Chem., 68, 9819–9820. (c) Mahoney, J.M., Marshall, R.A., Beatty, A.M., Smith, B.D., Camiolo, S. and Gale, P.A. (2001) J. Supramol. Chem., 1, 289–292. 27 Koulov, A.V., Mahoney, J.M., Beatty, A.M. and Smith, B.D. (2003) Org. Biomol. Chem., 1, 27–29. 28 Mahoney, J.M., Nawaratna, G.U., Beatty, A.M., Duggan, P.J. and Smith, B.D. (2004) Inorg. Chem., 43, 5902–5907. 29 Mahoney, J.M., Beatty, A.M. and Smith, B.D. (2004) Inorg. Chem., 43, 7617–7621. 30 Mahoney, J.M., Stucker, K.A., Jiang, H., Carmichael, I., Brinkmann, N.R.,

31

32

33 34 35

36 37

38

39 40

41

42

43

44 45

Beatty, A.M., Noll, B.C. and Smith, B.D. (2005) J. Am. Chem. Soc., 127, 2922–2928. Gunning, P., Benniston, A.C. and Peacock, R.D. (2004) Chem. Commun., 2226–2227. Benniston, A.C., Gunning, P. and Peacock, R.D. (2005) J. Org. Chem., 70, 115–123. Gunning, P.T. (2005) Org. Biomol. Chem., 3, 3877–3879. Jencks, W.P. (1981) Proc. Natl. Acad. Sci. U.S.A., 78, 4046–4050. Suksai, C., Leeladee, P., Jainuknan, D., Tuntulani, T., Muangsin, N., Chailapakul, O., Kongsaeree, P. and Pakavatchai, C. (2005) Terahedron. Lett., 46, 2765–2769. Mathews, S.E. and Beer, P.D. (2005) Supramol. Chem., 17, 411–435. Lhotak, P. (2005) Anion Receptors Based on Calixarenes’’ in Anion Sensing, (Topics in Current Chemistry) (ed. Stibor, J.), Springer, Heidelberg, 255, 65–95. Morzherin, Y., Rudkevich, D.M., Verboom, W. and Reinhoudt, D.N. (1993) J. Org.Chem., 58, 7602–7605. Cameron, B.R. and Loeb, S.J. (1997) Chem. Commun., 573–574. Sansone, F., Baldini, L., Casnati, A., Lazzarotto, M., Ugozzoli, F. and Ungaro, R. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 4842–4844. Miao, R., Zheng, Q.-Y., Chen, C.-F. and Huang, Z.-T. (2005) Tetrahedron Lett., 46, 2155–2158. Casnati, A., Pirondini, L., Pelizzi, N. and Ungaro, R. (2000) Supramol. Chem., 12, 53–65. Casnati, A., Fochi, M., Minari, P., Pochini, A., Reggiani, M. and Ungaro, R. (1996) Gazz. Chim. Ital., 126, 99–102. Gale, P.A. (1998) Tetrahedron Lett., 39, 3873–3876. (a) Camiolo, S., Gale, P.A., Ogden, M.I., Skelton, B.W. and White, A.H. (2001) J. Chem. Soc. Perkin Trans., 2, 1294–1298. (b) Camiolo, S., Gale,

References

46

47 48

49

50

51

52

53

54

55

56 57 58

59

60 61

P.A., Light, M.E. and Hursthouse, M.B. (2001) Supramol. Chem., 13, 613–615. Kim, S.K., Kang, B.-G., Koh, H.S., Yoon, Y.J., Jung, S.J., Jeong, B., Lee, K.-D. and Yoon, J. (2004) Org. Lett., 6, 4655–4658. Sebo, L., Schweizer, B. and Diederich, F. (2000) Helv. Chim. Acta, 83, 80–92. Qing, G.-Y., He, Y.-B., Zhao, Y., Hu, C.-C., Liu, S.-Y. and Yang, X. (2006) Eur. J. Org. Chem., 1574–1580. Yang, Y.S., Ko, S.W., Song, I.H., Ryu, B.J. and Nam, K.C. (2003) Bull. Korean Chem. Soc., 24, 681–683. Zlatuskova, P., Stibor, I., Tkadlecova, M. and Lhotak, P. (2004) Tetrahedron, 60, 11383–11390. Tumcharern, G., Tuntulani, T., Coles, S. J., Hursthouse, M.B. and Kilburn, J.D. (2003) Org. Lett., 5, 4971–4974. Kaewtong, C., Fuangswasdi, S., Muangsin, N., Chaichit, N., Vicens, J. and Pulpoka, B. (2006) Org. Lett., 8, 1561–1564. Nabeshima, T., Saiki, T., Iwabuchi, J. and Akine, S. (2005) J. Am. Chem. Soc., 127, 5507–5511. Lankshear, M.D., Cowley, A.R. and Beer, P.D. (2006) Chem. Commun., 612–614. Tongraung, P., Chantarasiri, N. and Tuntulani, T. (2003) Tetrahedron Lett., 44, 29–32. Sisson, A.L., Clare, J.P. and Davis, A.P. (2005) Chem. Commun., 5263–5265. Khatri, V.K., Upreti, S. and Pandey, P.S. (2006) Org. Lett., 8, 1755–1758. Davis, A.P., Gilmer, J.F. and Perry, J.J. (1996) Angew. Chem., 108, 1410–1413; Angew. Chem. Int. Ed., 35, 1312–1315. Bhattarai, K.B., Davis, A.P., Perry, J.J., Walter, C.J., Menzer, S. and Williams, D. J. (1997) J. Org. Chem., 62, 8463–8473. Camiolo, S. and Gale, P.A. (2000) Chem. Commun., 1129–1130. Woods, C.J., Camiolo, S., Light, M.E., Coles, S.J., Hursthouse, M.B., King, M. A., Gale, P.A. and Essex, J.W. (2002) J. Am. Chem. Soc., 124, 8644–8652.

j345

62 Anzenbacher, P., Jr. Jursíková, K., Lynch, V.M., Gale, P.A. and Sessler, J.L. (1999) J. Am. Chem. Soc., 121, 11020–11021. 63 Dukh, M., Drasar, P., Cerny, I., Pouzar, V., Shriver, J.A., Král, V. and Sessler, J.L. (2002) Supramol. Chem., 14, 237–244. 64 Miyaji, H., Anzenbacher, P., Jr. Sessler, J.L., Bleasdale, E.R. and Gale, P.A. (1999) Chem. Commun., 1723–1724. 65 Anzenbacher, P., Jr. Jursikova, K. and Sessler, J.L. (2000) J. Am. Chem. Soc., 122, 9350–9351. 66 Nishiyabu, R. and Anzenbacher, P. Jr. (2006) Org. Lett., 8, 359–362. 67 Sessler, J.L., Král, V., Shishkanova, T.V. and Gale, P.A. (2002) Proc. Natl. Acad. Sci. U.S.A., 99, 4848–4853. 68 Sessler, J.L., Gale, P.A. and Genge, J.W. (1998) Chem. Eur. J., 4, 1095–1099. 69 Anzenbacher, P., Jr. Nishiyabu, R. and Palacios, M.A. (2006) Coord. Chem. Rev., 250, 2929–2938. 70 Nishiyabu, R., Palacios, M.A., Dehaen, W. and Anzenbacher, P. Jr. (2006) J. Am. Chem. Soc., 128, 11496–11504. 71 Yoon, D.-W., Hwang, H. and Lee, C.-H. (2002) Angew. Chem., 114, 1835–1837; Angew. Chem. Int. Ed., 41, 1757–1759. 72 Lee, C.-H., Na, H.-K., Yoon, D.-W., Won, D.-H., Cho, W.-S., Lynch, V.M., Shevchuk, S.V. and Sessler, J.L. (2003) J. Am. Chem. Soc., 125, 7301–7306. 73 Lee, C.-H., Lee, J.-S., Na, H.-K., Yoon, D.-W., Miyaji, H., Cho, W.-S. and Sessler, J.L. (2005) J. Org. Chem., 70, 2067–2074. 74 Sessler, J.L., Cyr, M.J., Lynch, V., McGhee, E. and Ibers, J.A. (1990) J. Am. Chem. Soc., 112, 2810–2813. 75 Sessler, J.L., Cyr, M.J. and Burrell, A.K. (1991) SynLett, 127–134. 76 Shionoya, M., Furuta, H., Lynch, V., Harriman, A. and Sessler, J.L. (1992) J. Am. Chem. Soc., 114, 5714–5722. 77 Sessler, J.L., Cyr, M.J., Furuta, H., Král, V., Mody, T., Morishima, T., Shionoya, M. and Weghorn, S. (1993) Pure Appl. Chem., 65, 393–398.

346

j 9 Anion-binding Macrocycles 78 Scherer, M., Sessler, J.L., Gebauer, A. and Lynch, V. (1998) Chem. Commun., 85–86. 79 Furuta, H., Cyr, M.J. and Sessler, J.L. (1991) J. Am. Chem. Soc.,113, 6677–6678. 80 Iverson, B.L., Shreder, K., Král, V., Sansom, P., Lynch, V. and Sessler, J.L. (1996) J. Am. Chem. Soc.,118, 1608–1616. 81 Sessler, J.L., Furuta, H. and Král, V. (1993) Supramol. Chem., 1, 209–220. 82 Král, V. and Sessler, J.L. (1995) Tetrahedron, 51, 539–554. 83 Král, V., Andrievsky, A., Sessler, J.L. (1995) J. Chem. Soc., Chem. Commun., 2349–2351. 84 Cormode, D.P., Murray, S.S., Cowley, A. R. and Beer, P.D. (2006) Dalton Trans., 5135–5140. 85 Gong, J. and Gibb, B.C. (2005) Chem. Commun., 1393–1395. 86 Gong, J. and Gibb, B.C. (2005) Chem. Commun., 3319–3321. 87 Beer, P.D., Berry, N., Drew, M.G.B., Fox, O.D., Padilla-Tosta, M.E. and Patell, S., (2001) Chem. Commun., 199–200. 88 Berry, N.G., Pratt, M.D., Fox, O.D. and Beer, P.D. (2001) Supramol. Chem., 13, 677–682. 89 (a) Beer, P.D. and Wallace, W.W.H., (2005) Sensing, Templation and Self-Assembly by Macrocyclic Ligand Systems” in Macrocyclic Chemistry, Current Trends and Future Perspectives (ed. Gloe, K.) Springer, The Netherlands, pp.105–119. (b) Beer, P.D. and Bayly, S.R. (2005) Anion Sensing by Metal-Based Receptors”, in Topics in Current Chemistry, 255, Springer, Berlin, pp. 125–162. 90 Wong, W.W.H., Curiel, D., Cowley, A.R. and Beer, P.D. (2005) Dalton Trans., 359–364. 91 Webber, P.R.A., Drew, M.G.B., Hibbert, R. and Beer, P.D. (2004) Dalton Trans., 1127–1135. 92 Wong, W.W.H., Phipps, D.E. and Beer, P.D. (2004) Polyhedron, 2821–2829.

93 Beer, P.D., Berry, N.G., Cowley, A.R., Hayes, E.J., Oates, E.C. and Wong, W.W.H. (2003) Chem. Commun., 2408–2409. 94 Beer, P.D., Cheetham, A.G., Drew, M.G.B., Fox, O.D., Hayes, E.J. and Rolls, T.D. (2003) Dalton Trans., 603–611. 95 (a) Galbraith, S.G., Plieger, P.G. and Tasker, P.A. (2002) Chem. Commun., 2662–2663.(b) Miller, H.A., Laing, N., Parsons, S., Parkin, A., Tasker, P.A. and White, D.J. (2000) Dalton Trans., 3773–3782. 96 Plieger, P.G., Tasker, P.A. and Galbraith, S.G. (2004) Dalton Trans., 313–318. 97 Fujita, T. and Lehn, J.-M. (1988) Tetrahedron Lett., 29, 1709–1712. 98 Hossain, M.A., Llinares, J.M., Powell, D. and Bowman-James, K. (2001) Inorg. Chem., 40, 2936–2937. 99 Kang, S.O., Powell, D. and BowmanJames, K. (2005) J. Am. Chem. Soc., 127, 13478–13479. 100 Kang, S.O., Hossain, M.A., Powell, D. and Bowman-James, K. (2005) Chem. Commun., 328–330. 101 Kang, S.O., Powell, D., Day, V.W. and Bowman-James, K. (2006) Angew. Chem., 118, 1955–1959; Angew. Chem. Int. Ed., 45, 1921–1925. 102 Kubik, S., Kirchner, R., Nolting, D. and Seidel, J. (2002) J. Am. Chem. Soc., 124, 12752–12760. 103 Kubik, S., Goddard, R., Otto, S., Pohl, S., Reyheller, C. and Stuwe, S. (2005) Biosens. Bioelectron., 20, 2364–2375. 104 Otto, S. and Kubik, S. (2003) J. Am. Chem. Soc., 125, 7804–7805. 105 Brooks, S.J., Gale, P.A. and Light, M.E. (2006) Chem. Commun., 4344–4346. 106 Hirata, O., Takeuchi, M. and Shinkai, S. (2005) Chem. Commun., 3805–3807. 107 Francesconi, O., Ienco, A., Moneti, G., Nativi, C. and Roelens, S. (2006) Angew. Chem., 118, 6845–6848; Angew. Chem. Int. Ed., 46, 6693–6696. 108 Garcia-España, E., Diaz, P., Llinares, J.M. and Bianchi, A. (2006) Coord. Chem. Rev., 250, 2952–2986.

References

109 Miranda, C., Escarti, F., Lamarque, L., Yunta, M.J.R., Navarro, P., Garcia-España, E. and Jimeno, M.L. (2004) J. Am. Chem. Soc., 126, 823–833. 110 Sessler, J.L. and Seidel, D. (2003) Angew. Chem., 115, 5292–5333; Angew. Chem. Int. Ed., 42, 5134–5175. 111 Seidel, D., Lynch, V. and Sessler, J.L. (2002) Angew. Chem., 114, 1480–1483; Angew. Chem. Int. Ed., 41, 1422–1425. 112 Sessler, J.L., Katayev, E., Dan Pantos, G. and Ustynyuk, Y.A. (2004) Chem. Commun., 1276–1277. 113 Moyer, B.A. and Singh, R.P. (2004) Fundamentals and Applications of Anion Separations, Kluwer Academic/Plenum Publishers, New York. 114 Sessler, J.L., Katayev, E., Pantos, G.D., Scherbakov, P., Reshetova, M.D., Khrustalev, V.N., Lynch, V.M. and

115

116

117

118 119

j347

Ustynyuk, Y.A. (2005) J. Am. Chem. Soc., 127, 11442–11446. Katayev, E.A., Pantos, G.D., Reshetova, M.D., Khrustalev, V.N., Lynch, V.M., Ustynyuk, Y.A. and Sessler, J.L. (2005) Angew. Chem., 117, 7552–7556; Angew. Chem. Int. Ed., 44, 7386–7390. Sessler, J.L., Roznyatovskiy, V., Pantos, G.D., Borisova, N.E., Reshetova, M.D., Lynch, V.M., Khrustalev, V.N. and Ustynyuk, Y.A. (2005) Org. Lett., 7, 5277–5280. Fyfe, M.C.T., Glink, P.T., Menzer, S., Stoddart, J.F., White, A.J.P. and Williams, D.J. (1997) Angew. Chem., 109, 2158–2160; Angew. Chem., Int. Ed. Engl., 36, 2068–2070. Beer, P., Sambrook, M.R. and Curiel, D. (2006) Chem. Commun., 2105–2177. Ng, K.-Y., Cowley, A.R. and Beer, P.D. (2006) Chem. Commun., 3676–3678.

j349

10 Rotaxane and Catenane Synthesis James A. Wisner and Barry A. Blight

10.1 Introduction

The synthesis of interlocked molecules has become commonplace over the past 25 years with the gradual development of a number of highly facile template methods for their construction. What were once laboratory curiosities have now taken a prominent place in the broad ﬁeld of supramolecular chemistry, especially regarding their uses and further potential as molecular switches and machines [1]. We present here an overview of the main synthetic approaches to these molecules, with a focus on methods in which macrocyclization reactions result in interlocked products. The analysis is by no means meant to be comprehensive or exhaustive in detail, but rather to convey the variety and utility of the selected synthetic strategies in generating abiotic rotaxane and catenane superstructures. 10.1.1 General Comments

Interlocked molecules are those assemblies of two or more molecules which are linked by a so-called mechanical bond [2]. The individual molecules are not connected covalently in any way but are linked via their spatial relationship to one another. The nature of this phenomenon is such that it necessarily involves a macrocyclic component as one or more of the molecules which compose the assembly. The simplest forms of these assemblies are represented by a [2]rotaxane and a [2]catenane (Figure 10.1), where the bracketed numeral preceding the name indicates the number of individual molecules comprising the interlocked product. The synthesis of a rotaxane may be executed by a number of different routes, only one of which involves the formation of a macrocyclic component in the ﬁnal

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

350

j 10 Rotaxane and Catenane Synthesis

Figure 10.1 Clipping strategies in macrocyclization reactions resulting in interlocked products such as [2]rotaxanes and [2]catenanes.

interlocking step, which is commonly referred to as clipping. The synthesis of a catenane requires the formation of at least one of the two (or more) catenated macrocyclic rings in the ﬁnal step as a result of the topologically complex nature of the relationship between the rings. This gives rise to two different synthetic strategies for construction of a hypothetical [2]catenane, which entail either one macrocyclization step (strategy A) or two (strategy B) [3]. Interlocked structures may be synthesized in the absence of any preorganization of their precursors by statistical methods, but these are generally lower yielding and can be technically challenging [2b, 4]. The adoption of template approaches has greatly increased the efﬁciency, and often simplicity, of the construction of mechanically bonded species. However, the template employed and the synthetic step to enact the ring closure must be compatible. This requirement manifests itself in a couple of different considerations. First, and most obviously, the reaction conditions for ring closure must be tolerant of the functionality present in the components and not interfere with the relevant template interactions between them (i.e., break up the template). The templates used in many of the syntheses of interlocked products are often a combination of numerous weak non-covalent forces and, as such, can be highly sensitive to reaction conditions which are not considered harsh in normal covalent synthetic sequences (e.g., medium/ strong acids or bases, highly polar solvents). Second, the application of a template can be problematic in the case of macrocyclization reactions, which are often run under high-dilution conditions. If the template utilized is the result of an equilibrium complexation of the components, then high-dilution conditions are of little beneﬁt unless the association constants for the relevant equilibria lie sufﬁciently in favor of complex formation. Luckily, many of the template methods discussed below act to organize not only the precursors with respect to one another, but also to predispose (or localize) the pendant arms of the pre-macrocyclic building blocks for their intramolecular macrocyclization. Hence, the template can serve multiple purposes and obviate the need for high dilution to reduce the formation of oligomeric and higher order cyclic species, which are often undesirable side reactions.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

10.2 Macrocyclization Reactions Resulting in Interlocked Products 10.2.1 Williamson Ether Synthesis

Thereactionof oxo-anions with alkyl halidesorsulfonatesis of centralimportance in the synthesis of crown ethers, and thus holds a similar position in the synthetic development of supramolecular chemistry as a whole [5]. It is no surprise then that this was the macrocyclization reaction used in the synthesis of the ﬁrst metal-templated catenane by Sauvage and Dietrich-Buchecker in 1983 [6]. The authors employed a transition metal ion with a tetrahedral coordination sphere (Cuþ) to arrange two phenanthroline-based ligands (one macrocyclic (1) and one acyclic (2); strategy A) in an orthogonal and interpenetrated manner (Scheme 10.1). In the course of their investigations, they found that the polyether segment of the macrocyclic ligand must contain at least ﬁve ethyleneoxy units in order to allow efﬁcient interpenetration by the acyclic ligand [3]. The coordination complex formed is very stable, which is likely necessary in light of the reaction conditions used to convert it to the interlocked product. The ﬁnal catenate 3 (the -ate sufﬁx indicating its nature as a transition-metal complex) was formed in 27% yield by reaction of the phenolate end groups of the acyclic ligand with pentaethylene glycoldiiodide and further templated using cesium carbonate as the base. Removal of the Cuþ template to liberate the metal-free catenand 4 (the -and sufﬁx indicating the ligand character of the catenane) was achieved through treatment with tetramethylammonium cyanide. The same catenate can be synthesized in 27% yield by strategy B from a Cuþ complex of two acyclic ligands and two equivalents of the oligoethyleneoxydihalide [7]. Sauvage and coworkers have exploited these synthetic routes to produce a large number of catenanes using a variety of different dihalide linkers for the macrocyclization step, with typically similar yields [3]. An elaboration of this methodology also produced the ﬁrst molecular trefoil knot 5, which is a topologically complex macrocycle (Scheme 10.2a) [8]. The authors used a variation on strategy B to synthesize this product. In this case, rather than a single metal center, the authors began with the dinuclear double-helical template 6 (or helicate), which is intertwined and presents its terminal hydroxyl functions from the opposite ends of each of the two ligands which form the complex. Cyclization in this instance, using similar conditions to that for the catenane (strategy B), forms the knotted macrocycle in 3% yield. An even further extension is the synthesis of a doubly-intertwined [2]catenane 7 (or Solomon knot) which proceeds in 2% yield (strategy A) using a similar approach and trinuclear helicate 8 (Scheme 10.2b) [9]. The yields for both of these products can be greatly improved by using alkene metathesis for the ring closures instead [10,11]. 10.2.2 Quaternization of Aromatic Amines (Menschutkin Reaction)

In the early 1980s, Stoddart and coworkers investigated the complexation of pyridinium dications (e.g., paraquat) using crown ethers [12]. The complexation of

j351

Scheme 10.1 Cu+-templated catenane formation first reported by Sauvage and coworkers.

352

j 10 Rotaxane and Catenane Synthesis

Scheme 10.2 Structural representations of the trefoil knot (a) and doubly intertwined [2]catenane (b) by Sauvage and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j353

354

j 10 Rotaxane and Catenane Synthesis these organic cations occurs through a combination of p-p stacking, ion-dipole interactions, and C–H. . .O hydrogen bonding. An extension of this complexation to the synthesis of interlocked species involved the introduction of bifunctional bridging units such as a,a0 -dibromo-p-xylene, which have the potential to form tetracationic macrocyclic product 9 with 4,40 -dipyridyl [13]. The macrocycle (Scheme 10.3) can be synthesized from the dicationic U-shaped intermediate 10 and a,a0 -dibromo-p-xylene in the absence of any template in a yield of 12% by reﬂuxing for 18 h with MeCN as solvent. However, in the presence of a 3-molar excess of a diethyleneglycol-appended p-electron-rich hydroquinone template (11) and by running the reaction at room temperature for 9 days, the yield of macrocyclic product is improved to 35%; a change of solvent to DMF and the application of ultrahigh pressure enhanced the yield to 62% [14]. The macrocyclization reaction described above has been used to generate a great number of catenane (12) and rotaxane (13) architectures (Scheme 10.4) using both crown ethers [preformed macrocyclic components 14 (strategy A)] and hydroquinone-based dumbbell-shaped polyethers [preformed acyclic components 15 (clipping)] as templates [14b, 15]. These templates are also relatively robust with regard to the substitution of different groups into both the tetracationic cyclophane and the neutral frameworks. A representative selection of [2]catenanes synthesized by this route is given in Table 10.1[14b, 15a, 16]. It is remarkable to note that the ﬁrst [2]catenane synthesized (entry 2, Table 10.1) could be isolated in 74% yield [14,16]. The ﬁnal entry in the table is an example of an expanded version of the tetracationic macrocycle which incorporates a biphenylene spacer between the paraquat residues and can accommodate two p-stacked aryl rings within its cavity. The reaction results in the isolation of the [3]catenane product in 25% yield with very little (2%) [2]catenane isolated [16p]. A spectacular demonstration of the power of this approach by Stoddart and coworkers was the iterative synthesis of a [5] catenane which the authors dubbed Olympiadane after its resemblance to the Olympic rings [17]. The application of this template to the clipping of acyclic dumbbell-shaped molecules to form rotaxanes [14] has received less intensive scrutiny but involves a similar synthetic scheme and conditions. A few representative examples of [2]rotaxanes synthesized in this manner (16–20) are shown in Figure 10.2 [15c, 18–21]. The use of a 1,5-dihydroxynaphthalene core in these cases usually results in higher yields of the desired interlocked products [18,21]. 10.2.3 Condensation of Amines with Acid Chlorides

The ﬁrst examples of [2]catenanes templated by hydrogen bonding were reported in a similar time frame by both Hunter [22] and Vögtle and coworkers [23] in 1992 using nearly identical components in their syntheses (Scheme 10.5). Vögtle and coworkers simply condensed two equivalents of the diamine 21 (one equivalent acting as base) with one equivalent of 5-methoxyisophthaloyldichloride using

Scheme 10.3 Template-directed synthesis of a tetracationic cyclophane by Stoddart and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j355

Scheme 10.4 Macrocyclization in the synthesis of a [2]catenane and [2]rotaxane reported by Stoddart and coworkers.

356

j 10 Rotaxane and Catenane Synthesis

O

O

O

O

Table image 10.1

O

O

C

N

B

O

O

O

A(X), X = O,S

X O

O

O

O

L1

S

S

F

E

D

S

S

R2

R1

L2

+

I

H

G

L3

Br R4

+

R3

Br

L4

N

L

O

K

J Ln M

S

N

F F

F F N

M

OH OH N

N

N

Q

P

O

N

N

N

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j357

358

j 10 Rotaxane and Catenane Synthesis Table 10.1 Selected examples of catenane formation via

quaternization of aromatic amines. L1

L2

R1

R2

L3

L4

R3

R4

E

E

A(O)

A(O)

O

O

E

E

E

E

A(O)

A(O)

O

O

F

F

G H E

G E E

A(O) A(O) A(O)

A(O) A(O) A(O)

O O P

O O P

E E E

E E E

G N N C E E E G E E M E E D E

G N E C E E E G E E E E E G E

A(O) A(O) A(O) A(O) A(O) B A(S) A(S) A(O) A(O) A(O) A(O) A(O) A(O) A(O)

A(O) A(O) A(O) A(O) A(O) B A(S) A(S) A(O) A(O) A(O) A(O) A(O) A(O) A(O)

P O O O O O O O O O O Q Q O O

P O O O O O O O O O O O Q O O

E E E E J E E E L K M E E E I

E E E E J E E E L E E E E E I

Solvent MeCN DMF MeCN DMF DMF MeCN MeCN DMF DMF MeCN MeCN MeCN MeCN MeCN DMF DMF MeCN MeCN MeCN MeCN DMF DMF MeCN

Time

Pressure

Yield (%)

Ref.

2d 3d – 3d 10d 4d 7d 15 15d 5d 5d 5d 7d –b 4d 4d 19d 14d 14d 14d 14d 5d 7d

– 10 kbar –a 10 kbar –a –a –a –a –a –a –a –a –a –a –a –a –a –a –a –a –a –a –a

70 74 0b 28 51 59 19 38 51 0b 60 31 59 40 70 77 30 30 24 52 73 23 25c,2d

14b, 15a 16a 14b, 16a 16a 16a 16b 16c 16d 16d 16d 16e 16f 16g 16h 16i 16j 16j 16k 16l 16m 16n 16n 16o 16p

a

a

ambient pressure; no reaction time reported; c [3]catenane; d [2]catenane. b

catalytic DMAP and high dilution in chlorobenzene. This gave the [2]catenane 22 (R ¼ OMe) in 12% yield. Hunter, in an attempt to improve on the yield of the parent macrocycle he had synthesized previously, used a two-step approach wherein three quarters of the ﬁnal macrocycle was synthesized ﬁrst (23), reducing the reaction to a 1 þ 1 rather than a 2 þ 2 macrocyclization. This simpliﬁcation, and the use of NEt3 as the base in a methylene chloride solvent system, resulted in isolation of the [2]catenane 24 (R ¼ H) in 34% yield. Vögtle and coworkers have explored this system further using Hunters 1 þ 1 approach and different dicarbonylchloride connecting groups (e.g., 2,5-furandicarbonyldichloride), postulating a mechanism for the catenation based on the experimental data [24]. The authors posited that catenation likely involves formation of an initial macrocycle and subsequent threading of a partially reacted mixed amide/acid chloride intermediate through the macrocyclic cavity. Further intramolecular condensation of the threaded intermediate would provide the interlocked product. Leigh and coworkers reported the synthesis of a related [2]catenane system in 1995 based on the condensation of isophthaloyldichloride and para-xylylenediamine

Figure 10.2 Representative examples of various clipped [2]rotaxanes synthesized by Stoddart and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j359

Scheme 10.5 Hydrogen bond-templated self-assembly of [2]catenanes reported by both Vögtle and coworkers and Hunter and coworkers.

360

j 10 Rotaxane and Catenane Synthesis

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Scheme 10.6 Self-assembly of the hydrogen bond-directed [2]catenane formation by Leigh and coworkers.

in anhydrous chloroform with NEt3 as base (Scheme 10.6) [25]. The reaction is exceedingly simple to execute, and [2]catenane 25 can be isolated without the aid of chromatography in 20% yield. The generality of the template was explored further using a variety of dicarbonylchloride and diamine starting materials to form a diverse array of [2]catenanes (Table 10.2) in yields ranging from 18 to 27% [26]. The authors noted that the ideal reaction concentration for catenane formation in the case of 25 was 0.02 M at 298 K in chloroform. The use of more polar solvents (e.g., DMF or acetone) in all these cases gives rise to only macrocyclic and linear products, presumably because of disruption of the hydrogen bonding template. The mechanism for catenane formation in this system has been examined theoretically, and the results support not only the hydrogen-bonding basis for templation of the catenane but also templation of the macrocycle itself from intermediate linear oligomers [27]. The macrocyclization reaction discussed above can be adapted to the formation of more constitutionally varied interlocked structures using a clipping methodology. A number of different linear subunits with sterically demanding termini (26–31) have been used to template the clipping of macrocycle 32 to give [2]rotaxane products (Figure. 10.3) [28]. In these schemes, the template is the limiting reagent and the pre-macrocyclic starting materials are added in a 4- to 5-fold excess to maximize the yield of interlocked species. The fumaramide-based thread 29 templates the corresponding [2]rotaxane in 97% yield, which is astonishing given that the reaction sequence involves the condensation of four components encircling a ﬁfth [28–31]. A similar approach with preformed macrocyclic species that incorporate the same subunits yields catenated architectures with structurally dissimilar rings [32]. 10.2.4 Oxidative Acetylide Coupling

The oxidative coupling of two terminal alkynes via their copper acetylide complexes to form a 1,3-butadiyne function is known variously as Glaser, Eglinton, or Hay

j361

27

18

19

20

21

19

N

O

O

F

a

F

F

poor solubility for these products hampered all characterization attempts.

–

25

25

S

20

18

22

O2 N

F

20

20

Br

O

R

18

t-Bu

S

20

CH2Cl2

5

5

–

–

6

trace

6

5

–

–

7

–

2

5

–

trace

10

–a

(4þ4) Macrocycle [%]

trace

–a

(3þ3) Macrocycle [%]

(2+2) Macrocycle (2+2+2+2) Catenane (4+4) Macrocycle (4+4+2+2) Catenane

18

NH2

NEt3

25

H2 N

R

20

+

–a

Cl

O

(2þ2þ2þ2) Catenane [%]

Cl

S

(2þ2) Macrocycle [%]

O

Table 10.2 Representative examples of [2]catenanes synthesized by Leigh and coworkers

–

–

–

trace

–

–

trace

–

5

–

(4þ4þ2þ2) Catenane [%]

362

j 10 Rotaxane and Catenane Synthesis

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Figure 10.3 Examples of [2]rotaxane formation via hydrogen bond-templated macrocyclization reported by Leigh and coworkers.

coupling depending on the conditions used in the synthesis [33]. The reaction is highly facile and is a method of choice when constructing macrocyclic species containing 1,3-butadiynes in the backbone [34]. Given the wide precedent for this type of macrocyclization, the application of this methodology is a natural extension in the synthesis of interlocked species. Sauvage and coworkers have explored Glaser coupling in conjunction with their Cu+ template approach to synthesize multiply catenated structures [35]. The authors used a macrocyclic Cu+ complex 33 pre-threaded with acyclic bis-propargyl-terminated phenanthroline co-ligand 34 as the starting material (Scheme 10.7). Oxidative coupling of this complex (pathway A) in the presence of a very large excess of CuCl and CuCl2 produced a mixture of [3]- (35), [4]-, [5]-, [6]-, and [7]-catenate products. Characterization in this case was hampered by similar or identical NMR spectra and difﬁculty

j363

364

j 10 Rotaxane and Catenane Synthesis

Scheme 10.7 Oxidative acetylide coupling in the formation of [3]catenanes 35 and 39 reported by Sauvage and coworkers.

in separation of the higher order [5]-, [6]- and [7]catenates. A further investigation of a related system executed in a stepwise manner (i.e., intermolecular coupling (36, 37) followed by intramolecular coupling of the threaded macrocycles 38 (pathway B) resulted in an improved yield of [3]catenane 39 (40%) for the ﬁnal ring closure [36]. However, the synthetic difﬁculty in applying Cu+ as both template and reagent was apparently a limiting factor. Siegel and coworkers have recently applied a similar metal template approach with Ru(II), terpyridyl- (40) and pyridyl/phenanthroline-based (41) ligands using 20 equivalents of Cu(OAc)2/MeCN to enact a double ring closure (42–45) in the ﬁnal step, with yields of [2]catenates (46, 47) ranging from 20 to 25% (Scheme 10.8) [37].

Scheme 10.8 Synthesis of a ruthenium-templated [2]catenane by Siegel and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j365

366

j 10 Rotaxane and Catenane Synthesis Bäuerle and coworkers have adapted the Cuþ template (48) approach to [2]catenane synthesis using an intermediate platinum diacetylide linkage to macrocyclize each of the two rings (49, Scheme 10.9) [38]. Oxidation of the platinum centers in the macrocyclic rings of 50 with I2 induces reductive elimination of the two acetylides to form the 1,3-butadiyne linked macrocycles (51). Unfortunately, the authors could not remove the copper template in this example, likely as a result of steric congestion about the metal ion in the interlocked product. In 1997, Hamilton, Sanders, and coworkers introduced a new template for [2]catenane formation based solely on p-stacking of neutral donor and acceptor p-systems (Scheme 10.10) [39]. The authors used large excesses of CuCl/CuCl2 in DMF at 50 C to cyclize bifunctional pyromellitimide (52) and naphthalene diimide (53) monomers (2 equivalents/30 mM) in the presence of a dinaphtho-crown ether (1 equivalent) to form the corresponding [2]catenanes 54–55 and 56 in 25–33% and 29% yields, respectively. A later report in which the reactions were performed at room temperature, in spite of partial insolubility of the starting materials, raised the yields to 38% and 52%, respectively [40]. It is interesting to note that the mixed acceptor [2]catenane 56 (i.e., pyromellitimide/naphthodiimide) can be synthesized in 24% yield by reaction of equimolar amounts of all three starting materials. A remarkable extension of this methodology to the tandem closure of both catenated rings using the same acetylide coupling reaction, shown in Scheme 10.11, can be achieved in the case of the pyromellitimide acceptor. Cyclization of this starting material with acyclic propargyl-terminated dinaphthyl polyether 57 produces the corresponding [2]catenane 58 in 13–16% yield in one step [41]. The same reaction run with the naphthalenediimide starting material 53 failed. 10.2.5 Alkene Metathesis

The great impact that the development of alkene metathesis has had on synthetic chemistry [42] has likewise inﬂuenced the construction of interlocked molecules. The ready availability of catalyst, high yields, mild conditions and broad functional group tolerance of the Ring Closing Metathesis (RCM) version of this reaction has prompted its use in a wide range of template approaches for rotaxane and catenane formation. The thermodynamic distribution of the products obtained from the reaction as a result of its reversible nature can lead to very high yields of mechanically bonded products in a well-designed system. A complicating factor in many of these syntheses is the presence of mixtures of E/Z isomers in the disubstituted alkene products. The subsequent hydrogenation of the alkene is often used to remove the stereoisomerism and simplify the ﬁnal puriﬁcation of the interlocked species. A number of metal-templated catenates have been prepared by RCM in conjunction with a variety of 4-, 5-, and 6-coordinate metal templates using homogeneous and mixed-ligand systems. Several representative examples are given in Figure 10.4 highlighting the ﬁnal alkene connection used to cyclize the ring(s). The majority of these systems follow the general approach of Sauvage and Dietrich-Bucheckers

Scheme 10.9 Formation of a Pt(II)-linked, followed by a 1,3butadiyne-linked [2]catenane reported by Bäuerle and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j367

Scheme 10.10 Formation of p-donor/acceptor-templated [2]catenanes reported by Sanders and coworkers.

368

j 10 Rotaxane and Catenane Synthesis

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Scheme 10.11 Tandem macrocyclization and catenation via oxidative acetylide coupling reported by Sanders and coworkers.

original 3-D organization of the ligands with the metal template, which orients the pendant arms of the acyclic ligand(s) for macrocyclization (59–62) [43–46]. A subtly different method was applied by Leigh and coworkers. The authors employed an ostensibly 2-D square-planar palladium template and took advantage of electronic and steric inﬂuences in the intermediate metal complex to dispose the ligands appropriately (i.e., orthogonally) for formation of the ﬁnal interlocked products (63) [47]. Sanders and coworkers have used RCM as the ﬁnal ring-forming step in their neutral donor/acceptor p-stacking templated catenanes (64 and 65) with both pyromellitimide (66) and naphthodiimide (67) acceptors (Scheme 10.12) [48]. The addition of LiI during synthesis of the pyromellitimide-containing [2]catenane (64) increased the magnitude of the template interaction and resulted in a tripling of the yield to 60% [49]. X-ray crystallographic data of the corresponding [2]pseudorotaxane revealed the complexation of two Li cations by the oxygen donors of both the organic components. The well-established secondary dialkylammonium/crown ether template introduced by Stoddart and coworkers in 1995 for the formation of interlocked molecules has been adapted for RCM to form both rotaxanes and catenanes (Scheme 10.13). Mechanically bonded products (68, 69) have been synthesized by the clipping of either crown ether (70) around the ammonium dumbbell 71 or of the dialkylammonium component 72 around the crown ether 73 [50,51]. Higher yields were usually observed when the corresponding pre-formed macrocyclic alkenes were equilibrated with the other component of the template. Leigh and coworkers have demonstrated the adaptation of their hydrogen-bonded template methodology to RCM (Scheme 10.14), which can be used to reversibly form macrocycle 74 or [2]catenane 75 in high yields (>95%), depending on the synthetic conditions [52]. Böhmer and coworkers employed the homo- and heterodimerization of tetraurea-functionalized calix[4]arenes (76–78) and RCM to synthesize [2]rotaxanes (79) and [2]catenanes (80) which have multiply interlocked topologies (Scheme 10.15) [53,54]. In these cases, the authors enact multiple ring closures at high dilution,

j369

Figure 10.4 Selected examples of molecular architectures rendered interlocked via RCM.

370

j 10 Rotaxane and Catenane Synthesis

Scheme 10.12 RCM in the synthesis of [2]catenanes using the p-donor/acceptor approach by Sanders and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j371

372

j 10 Rotaxane and Catenane Synthesis

Scheme 10.13 Dialkylammonium and crown ether cyclizations via alkene metathesis.

Scheme 10.14 Reversible (concentration-dependent) RCM templated by hydrogen bonding.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Scheme 10.15 Synthesis of multiply interlocked molecular architectures reported by Böhmer and coworkers.

followed by hydrogenation, to give the ﬁnal products in 50–88% yield, depending on the particular substrates and number of connections. The recently developed anion template methodology developed by Beer and coworkers (Scheme 10.16) has been applied in conjunction with RCM to synthesize both [2]rotaxanes (81) and catenanes (82) from their respective starting materials (83–85) [55,56]. In this case, the two organic components are organized orthogonally around

j373

374

j 10 Rotaxane and Catenane Synthesis

Scheme 10.16 Anion-templated [2]rotaxane and [2]catenane formation reported by Beer and coworkers.

the halide using a combination of ion pairing and hydrogen bonding. The pendant arms of the acyclic component(s) are designed to interact non-covalently with the opposing ligand, which further organizes them for ring closure. In the absence of the halide template (e.g., employing only PF6– salts), little or no ring-closed product is generally observed. 10.2.6 Imine Formation/Reductive Amination

The formation of a Schiff base through condensation of a primary amine with an aldehyde or ketone has recently found application in the construction of

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Scheme 10.17 Synthesis of 87 via imine formation reported by Leigh and coworkers.

interlocked molecules [57]. The dynamic, reversible nature of such a connection lends itself well to template synthesis, in which the thermodynamically most stable product is desired and selected through supramolecular design of the components. The resulting equilibrium distribution of species in this case may be kinetically ﬁxed by reduction of the imine functions to their corresponding secondary amines. Leigh and coworkers have modiﬁed their octahedral metal-templated approach (62) to form [2]catenates through imine condensation (Scheme 10.17) [46]. Instead of using RCM (Figure 10.4) to cyclize the intermediate acyclic complex formed from an existing 2,6-diiminopyridine-based ligand, the authors condensed the corresponding diamine 86 with 2,6-diformylpyridine in situ with the octahedral metal (Zn2þ) to give [2]catenate 62 in 53% yield. Though the authors state that the yields for the other metal-templated catenates by this route are uniformally higher than those following the RCM methodology, the yields were not optimized, and a follow-up report has not yet been published. The iminecontaining Zn2þ [2]catenate synthesized by these methods was reduced with NaBH4 in EtOH and demetalated with EDTA to produce the related [2]catenand 87 in 78% yield. Nitschke and coworkers have self-assembled a [2]catenate (Scheme 10.18) [58] using a dicopper bis(phenanthrolinediimine) helicate subunit originally reported by Ziessel and coworkers [59]. Addition of an appropriately composed diamine (88) to 2,9- diformylphenanthroline and Cu(MeCN)4BF4 starting materials produced the [2]catenate 89 in quantitative yield. The corresponding [2]catenand 90 was generated by reduction (NaBH4) and demetalation (EDTA), although the authors did not purify the product because of the small scale of the reaction. The product was identiﬁed by MS/MS analysis and the yield was estimated at >50%.

j375

376

j 10 Rotaxane and Catenane Synthesis

Scheme 10.18 Self-assembly of a Cuþ-templated [2]catenane reported by Nitschke and coworkers.

Stoddart and coworkers have used imine condensation in two different contexts to form interlocked molecules. The authors have modiﬁed the crown ether component of their dialkylammonium/DB24C8 system to incorporate reversible imine subunits which allow for dynamic clipping around dialkylammonium cation 91 [60]. Three diformyl spacers (92–94) have been investigated with the same acyclic diamine 95 for their thermodynamic stability and kinetics of macrocyclization (96). However, reduction with BH3 to the kinetically trapped macrocyclic secondary diamine has only been reported for reaction of the 2,6-diformylpyridine derivative (Scheme 10.19; 97) [60a]. A spectacular demonstration of the power of imine condensation for the formation of interlocked species was recently revealed by Stoddart and coworkers. The authors reported the ﬁrst synthesis of a molecular version of the Borromean Rings, in which three rings are interlocked such that scission of any one ring releases all three (Scheme 10.20) [61]. The synthesis of this product is even more remarkable given the method used to generate it. The ﬁnal interlocked product is self-assembled through condensation of six diamine (98), six 2,6-diformylpyridine, and six divalent transition metal components in a single synthetic step. The conversions vary in yield from a low of 36% with Co2þ to an astonishing 95% in the case of Zn2þ templation (99). The Zn2þ-templated Borromeate 99 has been further reduced (NaBH4) and demetalated (EDTA), resulting in the fully interlocked (i.e., non-dynamic) Borromean link topology (100) [62]. It is interesting to note that the use of a 1:1 mixture of metals (Cu2þ and Zn2þ) in the self-assembly reaction results in a different product. In this case, the

Scheme 10.19 Synthesis of various [2]rotaxanes by imine formation around a dialkylammonium dumbbell reported by Stoddart and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j377

Scheme 10.20 Synthetic molecular Borromean rings and a Solomon knot reported by Stoddart and coworkers.

378

j 10 Rotaxane and Catenane Synthesis

10.2 Macrocyclization Reactions Resulting in Interlocked Products

isolated material is in the form of a Solomon Knot or doubly intertwined [2]catenane (101) [63]. 10.2.7 Metal-Ligand Coordination

A particularly facile synthetic method in the context of mechanically bonded molecules is the closure of one or more of the rings in the ﬁnal products by using labile metal–ligand coordination as a linkage in the backbone under thermodynamic control. The underlying complication inherent in this approach is the very lability of the dative bonds used in the self-assembly of the macrocyclic component(s). Unless some method is available (e.g., solvent, temperature) to change the kinetic stability of the metal-ligand connection, the interlocked complex is forever at the mercy of equilibrium in solution. Hence, a change in the solution conditions (e.g., concentration) often changes the position of the equilibrium to give a mixture of interlocked and non-interlocked species. We will therefore focus here only on examples in which the mechanically bonded products have demonstrated kinetic stability and are isolable in the solution state without interference from further dissociation into their component parts. In 1994, Fujita and coworkers introduced a route to [2]catenanes from the self-association of two metallacyclic rings through hydrophobic interactions in aqueous solution (Scheme 10.21) [64]. The self-assembled Pd-based macrocycle 102, from ligand 103, is in equilibrium at room temperature between the catenated product 104 and the free component rings. In contrast, the analogous Pt macrocycle 105 is self-assembled in water at 100 C but does not form a [2]catenane. Cooling to room temperature locks the macrocyclic geometry kinetically [65]. The addition of 5 M NaNO3 and further heating to 100 C for a period of 24 h converted the macrocycle quantitatively into [2]catenane 106, which was veriﬁed upon cooling and precipitation as the perchlorate salt. The authors describe this system as a molecular lock, in that the catenane topology does not undergo dissociation even upon heating in water at 100 C. A similar approach has been used to self-assemble a catenated cage compound from ten separate components [66]. Very recently, Fujita and coworkers have extended this methodology to catenation in a related system at room temperature (Scheme 10.22) [67]. A solution of macrocycle 107 (2 mM) was found to be kinetically stable in 1:2 DMSO-d6/D2O with no catenated products observed. However, upon irradiation with UV light (Hg lamp; 330 70 nm band-pass ﬁlter) the starting macrocycle is converted quantitatively into the corresponding [2]catenane 108. Dissolution of the catenane in non-aqueous solvent (MeCN/MeOH 1:1), which would disassemble the catenane if ligand exchange were occurring, produced no change in the interlocked topology. Irradiation under these conditions disassembles the catenane into the original component rings. The authors also examined the analogous Ru-based approach, which demonstrated the same behavior. Sauvage, Fujita, and coworkers have combined both their copper template and palladium self-assembly methodologies to form both a doubly-intertwined

j379

Scheme 10.21 Reversible (104) and non-reversible (106) formation of metal-based [2]catenanes by Fujita and coworkers.

380

j 10 Rotaxane and Catenane Synthesis

10.2 Macrocyclization Reactions Resulting in Interlocked Products

Scheme 10.22 Photoswitchable [2]catenane formation reported by Fujita and coworkers.

[2]catenane or Solomon knot (109 from ligand 110) and a simple [2]catenane (111 from ligand 112) [68,69]. The doubly intertwined [2]catenane was demonstrated by the authors to be the ﬁnal products irrespective of the order in which the selfassembly and template steps were used in the synthetic sequence (Scheme 10.23). In this case, however, the Cuþ template cannot be removed. Jeong and coworkers have used metal-ligand coordination to form a metalomacrocyclic complex (113, Scheme 10.24) which closely resembles the macrocyclic component of 24 (Scheme 10.5) when cyclized [70]. The authors use dumbbellshaped axles, similar to those pioneered by Vögtle and coworkers [71], which act as templates about which 113 (M ¼ Re(CO)3Cl) may be clipped at elevated temperature. In the case of this Re-based macrocycle, the resulting [2]rotaxanes (114) are stable at room temperature and may be chromatographed on silica gel. Quintela, Peinador, and coworkers have adapted Stoddarts paraquat box to include metalligand coordination in the backbone by replacing one p-xylenediyl group with a 4picolyl subunit (Scheme 10.25) [72]. The corresponding [2]catenane 115 was formed in 85% yield by reﬂuxing the ligand (116), (en)Pt(NO3)2, and 14 at 100 C for 12 days. A subtly different approach has been taken by Wisner and coworkers using a method in which the same metal center acts as both the template and the connection point in the macrocyclic backbone to form a [2]catenane [73]. The template is based on the complexation of PdCl2L2 complexes by the macrocycle derived from Hunters [2]catenane synthesis [74]. In this case, the opposing isophthalamide subunits in the macrocycle hydrogen bond to the chloride ligands to form a [2]rotaxane or [2]pseudorotaxane complex [75,76]. The basic template can be adapted to form a [2]catenane by incorporating the PdCl2 subunit into the backbone of the macrocycle (ligand 117 þ PdCl2). The resulting metallacycle is capable of self-recognition, which results in a mutually interlocked catenane- architecture (118, Scheme 10.26). The [2]catenane is remarkably stable given the typical lability of the Pd-N coordinate bonds used to form each macrocycle. The catenane can be chromatographed if need be and is stable to a range of normally competitive solvents. However, addition of >1 M DMSO induces the dissociation of the two rings into the parent macrocycles. The catenated topology can be easily re-established in quantitative yield by dissolution in a less competitive solvent.

j381

Scheme 10.23 The formation of a Solomon Knot (109) and a [2]catenane (111) reported by Sauvage, Fujita, and coworkers.

382

j 10 Rotaxane and Catenane Synthesis

Scheme 10.24 [2]Rotaxane formation by metallomacrocyclic clipping reported by Jeong and coworkers.

10.2 Macrocyclization Reactions Resulting in Interlocked Products

j383

384

j 10 Rotaxane and Catenane Synthesis

Scheme 10.25 [2]Catenane formation by metallomacrocyclic clipping reported by Quintela and coworkers.

10.3 Conclusions

The use of macrocyclization reactions to form interlocked molecules has progressed signiﬁcantly over the last 25 years. The challenge of forming these mechanically bonded architectures has been tempered somewhat with the application of both an ever wider variety of synthetic methods and increasingly sophisticated templates. However, the exploitation of new synthetic techniques (e.g., alkene metathesis), a further understanding of the parameters which guide self-assembly (e.g., Borromean Rings), and a continuing shift from aesthetics to function will likely shape the future character of this ﬁeld. The truly unique opportunities that these types of molecules offer ensures that the exploration of their syntheses will remain an exciting area of endeavor for many years to come.

10.4 Experimental: Selected Procedures 10.4.1 [2]Catenane 14 [14b,16a]

A solution of 102PF6 (140 mg, 0.20 mmol) in anhydrous CH3CN (5.0 mL) was added to a stirred solution of 14 (280 mg, 0.52 mmol), also in anhydrous CH3CN (5.0 mL). A deep yellow color ensued, and a solution of 1,4-bis(bromomethyl)benzene (56 mg, 0.20 mmol) in anhydrous CH3CN (5.0 mL) was added. Upon stirring for 4 h, the solution turned a deep red color. The solution was stirred for a total of

Scheme 10.26 Metallomacrocyclic/[2]catenane formation templated by second-sphere coordination of the PdCl2 subunit reported by Wisner and coworkers.

10.4 Experimental: Selected Procedures

j385

386

j 10 Rotaxane and Catenane Synthesis 48 h at rt, and the resulting red precipitate was ﬁltered and washed with CHCl3 (10 mL), dissolved in H2O (30 mL), and ﬁltered to remove any insoluble material. An aqueous solution of NH4PF6 was added to the ﬁltrate until no further precipitation occurred. The precipitate was ﬁltered off, washed with H2O (10 mL), and dried. Recrystallization from CH3CN and H2O gave [2]catenane 12 (230 mg, 70% yield). 10.4.2 [2]Catenane 25 [25]

Isophthaloyldichloride (0.87 g, 4.3 mmol) and 1,4-xylylenediamine (0.58 g, 4.3 mmol), separately dissolved in anhydrous CHCl3 (130 mL each), were added via syringe pump over 30 min to a stirred solution of triethylamine (1.19 g, 18.9 mmol) in anhydrous CHCl3 (130 mL). Following the addition, the reaction was stirred for an additional 12 h. The precipitate was ﬁltered off, and the ﬁltrate was washed with 1 M aqueous HCl (3 · 200 mL), 5% aqueous NaOH (3 · 200 mL), and H2O (3 · 200 mL). It was then dried with MgSO4, and the volume was reduced in vacuo to afford 0.23 g (21%) of [2]Catenane 25. 10.4.3 [2]Rotaxane 81 [55]

Compound 83 (0.125 g, 0.21 mmol) and compound 84 (0.125 g, 0.12 mmol) were dissolved in anhydrous dichloromethane (5 mL), and stirred for 5 min. Grubbs catalyst (20 mol%) was then added to the solution, and the reaction mixture was stirred for an additional 16 h. Solvent was removed under reduced pressure, and the crude product chromatographed (CHCl3:CH3OH, 93:7), collecting the ﬁrst yellow band eluted. The product was recrystallized by slow diffusion of diisopropyl ether into chloroform to yield brilliant yellow crystals of [2]rotaxane 81 (0.091 g, 47%).

10.4.4 [2]Catenane 118 [73]

Dipyridyl ligand 117 (0.125 g, 0.142 mmol) was dissolved in a 3:1 CHCl3/CH3CN solvent mixture (10 mL). A chloroform solution (2 mL) of trans-bis-benzonitrile palladium(II)dichloride (0.0543 g, 0.142 mmol) was added dropwise to the reaction mixture, which was then reﬂuxed for 8 h. The crude reaction mixture was ﬁltered through celite to remove any precipitate, and this was followed by the removal of solvent under reduced pressure. The resulting solid was dissolved in a minimum amount of dichloromethane and precipitated with hexanes to remove residual benzonitrile, and the precipitate was ﬁltered to afford [2]catenane 118 (0.129 g, 87%).

References

Acknowledgments

We would like to acknowledge the continued support of our research by the Natural Science and Engineering Council of Canada and the University of Western Ontario.

Abbreviations

TIPS Py Terpy Phenpy dppp RCM EDTA Lut en TMS Grubbs Cat. 1

Triisopropylsilyl Pyridine 2,20 ,60 ,200 -Terpyridine 2-(Pyridin-2-yl)-1,10-phenanthroline 1,3-Bis(diphenylphosphino)propane Ring Closing Metathesis Ethylenediamine tetraacetic acid Lutidine Ethylenediamine Trimethylsilyl Generation 1 Grubbs Catalyst

References 1 Kay, E.R., Leigh, D.A. and Zerbetto, F. (2007) Angew. Chem., 119, 72–196; Angew. Chem. Int. Ed., 46, 72–191. 2 J.-P. Sauvage and C. Dietrich-Buchecker (eds) (1999) Molecular Catenanes, Rotaxanes and Knots: A Journey Through the World of Molecular Topology, WileyVCH, Weinheim; (b) Schill, G. (1971) Catenanes, Rotaxanes, and Knots, Academic Press, New York. 3 Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1987) Chem. Rev., 87, 795–810. 4 (a) Wasserman, E. (1960) J. Am. Chem. Soc., 82, 4433–4434; (b) Frisch, H.L. and Wasserman, E. (1961) J. Am. Chem. Soc., 83, 3789–3795; (c) Harrison, I.T. and Harrison, S. (1967) J. Am. Chem. Soc., 89, 5723–5724; (d) Harrison, I.T. (1972) J. Chem. Soc., Chem. Commun., 231–232; (e) Harrison, I.T. (1974) J. Chem. Soc., Perkin Trans. 1, 301–304; (f) Agam, G., Graiver, D. and Zilkha, A. (1976) J. Am.

5

6

7

8

9

Chem. Soc., 98, 5206–5214; (g) Agam, G. and Zilkha, A. (1976) J. Am. Chem. Soc., 98, 5214–5216. (a) Pedersen, C.J. (1967) J. Am. Chem. Soc., 89, 7017–7036; (b) G.W. Gokel (ed.) (1991) Crown Ethers and Cryptands, Royal Society of Chemistry, Cambridge. Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1983) Tetrahedron Lett., 24, 5095– 5098. Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1984) J. Am. Chem. Soc., 106, 3043– 3045. (a) Dietrich-Buchecker, C.O., Sauvage, J.-P. and Kern, J.-M. (1989) Angew. Chem., 101, 192–194; Angew. Chem. Int. Ed., 28, 189–192; (b) Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1992) New. J. Chem., 16, 931–942. (a) Nierengarten, J.-F., DietrichBuchecker, C.O. and Sauvage, J.-P. (1994) J. Am. Chem. Soc., 116, 375–376;

j387

388

j 10 Rotaxane and Catenane Synthesis

10

11 12

13

14

15

16

(b) Nierengarten, J.-F., DietrichBuchecker, C.O. and Sauvage, J.-P. (1996) New. J. Chem., 20, 685–693. Rapenne, G., Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1999) J. Am. Chem. Soc., 121, 994–1001. Dietrich-Buchecker, C.O. and Sauvage, J.-P. (1999) Chem. Commun., 615–616 (a) Allwood, B.L., Shahriari-Zavareh, H., Stoddart, J.F. and Williams, D.J. (1987) J. Chem. Soc. Chem. Commun., 1058– 1061; (b) Allwood, B.L., Spencer, N., Shahriari-Zavareh, H., Stoddart, J.F. and Williams, D.J. (1987) J. Chem. Soc. Chem. Commun., 1061–1064; (c) Allwood, B.L., Spencer, N., Shahriari-Zavareh, H., Stoddart, J.F. and Williams, D.J. (1987) J. Chem. Soc. Chem. Commun., 1064– 1066. (a) Odell, B., Reddington, M.V., Slawin, A.M.Z., Spencer, N., Stoddart, J.F. and Williams, D.J. (1988) Angew. Chem., 100, 1605–1608; (1988) Angew. Chem. Int. Ed., 27, 1547–1550. (a) Brown, C.L., Philp, D. and Stoddart, J.F. (1991) Synlett, 462–464; (b) Anelli, P. L., Ashton, P.R., Ballardini, R., Balzani, V., Delgado, M., Gandolﬁ, M.T., Goodnow, T.T., Kaifer, A.E., Philp, D., Pietraszkiewicz, M., Prodi, L., Reddington, M.V., Slawin, A.M.Z., Spencer, N., Stoddart, J.F., Vincent, C. and Williams, D.J. (1992) J. Am. Chem. Soc., 114, 193–218. (a) Brown, C.L., Philp, D. and Stoddart, J.F. (1991) Synlett, 459–461; (b) Philp, D. and Stoddart, J.F. (1991) Synlett., 445– 458; (c) Ashton, P.R., Grognuz, M., Slawin, A.M.Z., Stoddart, J.F. and Williams, D.J. (1991) Tetrahedron Lett., 32, 6235–6238; (d) Brown, C.L., Philp, D., Spencer, N. and Stoddart, J.F. (1992) Israel J. Chem., 32, 61–67. (a) Amabilino, D.B., Anelli, P.-L., Ashton, P.R., Brown, G.R., Córdova, E., Godínez, L.A., Hayes, W., Kaifer, A.E., Philp, D., Slawin, A.M.Z., Spencer, N., Stoddart, J.F., Tolley, M.S. and Williams, D.J. (1995) J. Am. Chem. Soc., 117,

11142–11170; (b) Asakawa, M., Ashton, P.R., Boyd, S.E., Brown, C.L., Gillard, R. E., Kocian, Ö., Raymo, F.M., Stoddart, J. F., Tolley, M.S., White, A.J.P. and Williams, D.J. (1997) J. Org. Chem., 62, 26–37; (c) Ballardini, R., Balzani, V., Credi, A., Gandolﬁ, M.T., Marquis, D., Pérez-García, L. and Stoddart, J.F. (1998) Eur. J. Org. Chem., 81–89; (d) Ashton, P.R., Ballardini, R., Balzani, V., Credi, A., Gandolﬁ, M.T., Menzer, S., Pérez-García, L., Prodi, L., Stoddart, J.F., Venturi, M., White, A.J.P. and Williams, D.J. (1995) J. Am. Chem. Soc., 117, 11171–11197; (e) Ballardini, R., Balzani, V., Credi, A., Brown, C.L., Gillard, R.E., Montalti, M., Philp, D., Stoddart, J.F., Venturi, M., White, A.J.P., Williams, B.J. and Williams, D.J. (1997) J. Am. Chem. Soc., 119, 12503–12513; (f) Gillard, R.E., Stoddart, J.F., White, A.J.P., Williams, B. J. and Williams, D.J. (1996) J. Org. Chem., 61, 4504–4505; (g) Ashton, P.R., Blower, M.A., McLean, C.H., Stoddart, J. F. and Tolley, M.S. (1994) Synlett, 1063– 1066; (h) Ashton, P.R., Preece, J.A., Stoddart, J.F., Tolley, M.S., White, A.J.P. and Williams, D.J. (1994) Synthesis, 1344–1352; (i) Ashton, P.R., Blower, M. A., Iqbal, S. McLean, C.H., Stoddart, J.F., Tolley, M.S. and Williams, D.J. (1994) Synlett, 1059–1062; (j) Asakawa, M., Ashton, P.R., Dehaen, W., L'abbe, G., Menzer, S., Nouwen, J., Raymo, F.M., Stoddart, J.F., Tolley, M.S., Toppet, S., White, A.J.P. and Williams, D.J. (1997) Chem. Eur. J., 3, 772–787; (k) Asakawa, M., Ashton, P.R., Brown, C.L., Fyfe, M. C.T., Menzer, S., Pasini, D., Scheuer, C., Spencer, N., Stoddart, J.F., White, A.J.P. and Williams, D.J. (1997) Chem. Eur. J., 3, 1136–1150; (l) Ashton, P.R., Balzani, V., Credi, A., Kocian, O., Pasini, D., Prodi, L., Spencer, N., Stoddart, J.F., Tolley, M.S., Venturi, M., White, A.J.P. and Williams, D.J. (1998) Chem. Eur. J., 4, 590–607; (m) Ashton, P.R., Heiss, A. M., Pasini, D., Raymo, F.M., Shipway, A. N., Stoddart, J.F. and Spencer, N. (1999)

References

17

18

19

20

21

22

23

Eur. J. Org. Chem., 995–1004; (n) Ashton, P.R., Boyd, S.E., Brindle, A., Langford, S. J., Menzer, S., Pérez-García, L., Preece, J. A., Raymo, F.M., Spencer, N., Stoddart, J.F., White, A.J.P. and Williams, D.J. (1999) New. J. Chem., 23, 587–602; (o) Asakawa, M., Ashton, P.R., Balzani, V., Credi, A., Hamers, C., Mattersteig, G., Montalti, M., Shipway, A.N., Spencer, N., Stoddart, J.F., Tolley, M.S., Venturi, M., White, A.J.P. and Williams, D.J. (1998) Angew. Chem., 110, 357–361; Angew. Chem. Int. Ed., 37, 333–337; (p) Amabilino, D.B., Ashton, P.R., Brown, C.L., Cordóva, E., Godínez, L.A., Goodnow, T.T., Kaifer, A.E., Newton, S. P., Pietraszkiewicz, M., Philp, D., Raymo, F.M., Reder, A.S., Rutland, M.T., Slawin, A.M.Z., Spencer, N., Stoddart, J. F. and Williams, D.J. (1995) J. Am. Chem. Soc., 117, 1271–1293. (a) Amabilino, D.B., Ashton, P.R., Reder, A.S., Spencer, N. and Stoddart, J.F. (1994) Angew. Chem., 106, 1316– 1319; Angew. Chem. Int. Ed., 33, 1286– 1290. Bravo, J.A., Raymo, F.M., Stoddart, J.F., White, A.J.P. and Williams, D.J. (1998) Eur. J. Org. Chem., 2565–2571 Anelli, P.L., Spencer, N. and Stoddart, J. F. (1991) J. Am. Chem. Soc., 113, 5131– 5133. Ballardini, R., Balzani, V., Dehaen, W., Dell'Erba, A.E., Raymo, F.M., Stoddart, J.F. and Venturi, M. (2000) Eur. J. Org. Chem., 591–602 Jeppesen, J.O., Perkins, J., Becher, J. and Stoddart, J.F. (2001) Angew. Chem., 113, 1256–1261; Angew. Chem. Int. Ed., 40, 1216–1221. (a) Hunter, C.A., (1992) J. Am. Chem. Soc., 114, 5303–5311; (b) Carver, F.J., Hunter, C.A. and Shannon, R.J. (1994) J. Chem. Soc. Chem. Commun., 1277– 1280; (c) Adams, H., Carver, F.J. and Hunter, C.A. (1995) J. Chem. Soc. Chem. Commun., 809–810. (a) Vögtle, F., Meier, S. and Hoss, R. (1992) Angew. Chem., 104, 1628–1631;

24

25

26

27

28

29

30

31

32

33

34 35

Angew. Chem. Int. Ed., 31, 1619–1622; (b) Ottens-Hildebrandt, S., Meier, S., Schmidt, W. and Vögtle, F. (1994) Angew. Chem., 106, 1818–1821; Angew. Chem. Int. Ed., 33, 1767–1770. Otten-Hildebrandt, S., Nieger, M., Rissanen, K., Rouvinen, J., Meier, S., Harder, G. and Vögtle, F. (1995) J. Chem. Soc. Chem. Commun., 777–778 Johnston, A.G., Leigh, D.A., Pritchard, R.J. and Deegan, M.D. (1995) Angew. Chem., 107, 1324–1327; Angew. Chem. Int. Ed., 34, 1209–1212. Johnston, A.G., Leigh, D.A., Nezhat, L., Smart, J.P. and Deegan, M.D. (1995) Angew. Chem., 107, 1327–1331; Angew. Chem. Int. Ed., 34, 1212–1216. Leigh, D.A., Venturini, A., Wilson, A. J., Wong, J.K.Y. and Zerbetto, F. (2004) Chem. Eur. J., 10, 4960–4969. Johnston, A.G., Leigh, D.A., Murphy, A., Smart, J.P. and Deegan, M.D. (1996) J. Am. Chem. Soc., 118, 10662–10663. Leigh, D.A., Murphy, A., Smart, J.P. and Slawin, A.M.Z. (1997) Angew. Chem., 109, 752–756; Angew. Chem. Int. Ed., 36, 728–732. Bermudez, V., Capron, N., Gase, T., Gatti, F.G., Kajzar, F., Leigh, D.A., Zerbetto, F. and Zhang, S. (2000) Nature, 406, 608–611. Gatti, F.G., Leigh, D.A., Nepogodiev, S. A., Slawin, A.M.Z., Teat, S.J. and Wong, J.K.Y. (2001) J. Am. Chem. Soc., 123, 5983–5989. (a) Leigh, D.A., Wong, J.K.Y., Dehez, F. and Zerbetto, F. (2003) Nature, 424, 174–179; (b) Hernández, J.V., Kay, E.R. and Leigh, D.A. (2004) Science, 306, 1532–1537. Siemsen, P., Livingston, R.C. and Diederich, F. (2000) Angew. Chem., 112, 2740–2767; Angew. Chem. Int. Ed., 39, 2632–2657. Maraval, V. and Chauvin, R. (2006) Chem. Rev., 106, 5317–5343. (a) Dietrich-Buchecker, C.O., Khemiss, A. and Sauvage, J.-P. (1986) J. Chem. Soc. Chem. Commun., 1376–1378;

j389

390

j 10 Rotaxane and Catenane Synthesis

36

37

38

39

40

41

42

43

44

(b) Dietrich-Buchecker, C.O., Guilhem, J., Khémiss, A.-K., Kintzinger, J.-P., Pascard, C. and Sauvage, J.-P. (1987) Angew. Chem., 99, 711–714; Angew. Chem. Int. Ed., 26, 661–663; (c) Sauvage, J.-P., (1990) Acc. Chem. Res., 23, 319–327; (d) DietrichBuchecker, C.O., Hemmert, C., Khémiss, A.-K. and Sauvage, J.-P. (1990) J. Am. Chem. Soc., 112, 8002–8008; (e) Bitsch, F., Dietrich-Buchecker, C.O., Khémiss, A.K., Sauvage, J.-P. and Van Dorsselaer, A. (1991) J. Am. Chem. Soc., 113, 4023–4025. Kern, J.-M., Sauvage, J.-P. and Weidmann, J.-L. (1996) Tetrahedron, 52, 10921–10934. Loren, J.C., Gantzel, P., Linden, A. and Siegel, J.S. (2005) Org. Biomol. Chem., 3, 3105–3116. Ammann, M., Rang, A., Schalley, C.A. and Bäuerle, P. (2006) Eur. J. Org. Chem., 1940–1948. Hamilton, D.G., Sanders, J.K.M., Davies, J.E., Clegg, W. and Teat, S.J. (1997) Chem. Commun., 897–898. Hamilton, D.G., Davies, J.E., Prodi, L. and Sanders, J.K.M. (1998) Chem. Eur. J., 4, 608–620. (a) Hamilton, D.G., Feeder, N., Prodi, L., Teat, S.J., Clegg, W. and Sanders, J.K.M. (1998) J. Am. Chem. Soc., 120, 1096– 1097; (b) Hamilton, D.G., Prodi, L., Feeder, N. and Sanders, J.K.M. (1999) J. Chem. Soc. Perkin. Trans. 1, 1057–1065. (a) Chauvin, Y. (2006) Angew. Chem., 118, 3824–3831; Angew. Chem. Int. Ed., 45, 3740–3747; (b) Schrock, R.R., (2006) Angew. Chem., 118, 3832–3844; Angew. Chem. Int. Ed., 45, 3748–3759; (c) Grubbs, R.H., (2006) Angew. Chem., 118, 3845–3850; Angew. Chem. Int. Ed., 45, 3760–3765. (a) Mohr, B., Weck, M., Sauvage, J.-P. and Grubbs, R.H. (1997) Angew. Chem., 109, 1365–1367; Angew. Chem. Int. Ed., 36, 1308–1310; (b) Weck, M., Mohr, B., Sauvage, J.-P. and Grubbs, R.H. (1999) J. Org. Chem., 64, 5463–5471. Hamann, C., Kern, J.-M. and Sauvage, J.-P. (2003) Inorg. Chem., 42, 1877–1883.

45 (a) Mobian, P., Kern, J.-M. and Sauvage, J.-P. (2003) J. Am. Chem. Soc., 125, 2016– 2017; (b) Mobian, P., Kern, J.-M. and Sauvage, J.-P. (2003) Inorg. Chem., 42, 8633–8637; (c) Arico, F., Mobian, P., Kern, J.-M. and Sauvage, J.-P. (2003) Org. Lett., 5, 1887–1890. 46 Leigh, D.A., Lusby, P.J., Teat, S.J., Wilson, A.J. and Wong, J.K.Y. (2001) Angew. Chem., 113, 1586–1591; Angew. Chem. Int. Ed., 40, 1538–1543. 47 (a) Fuller, A.-M., Leigh, D.A., Lusby, P.J., Oswald, I.D.H., Parsons, S. and Walker, D.B. (2004) Angew. Chem., 116, 4004– 4008; Angew. Chem. Int. Ed., 43, 3914– 3918; (b) Fuller, A.-M., Leigh, D.A., Lusby, P.J., Slawin, A.M.Z. and Walker, D.B. (2005) J. Am. Chem. Soc., 127, 12612–12619. 48 Hamilton, D.G., Feeder, N., Teat, S.J. and Sanders, J.K.M. (1998) New J. Chem., 1019–1021 49 Kaiser, G., Jarrosson, T., Otto, S., Ng, Y.-F., Bond, A.D. and Sanders, J.K.M. (2004) Angew. Chem., 116, 1993–1996; Angew. Chem. Int. Ed., 43, 1959–1962. 50 (a) Kilbinger, A.F.M., Cantrill, S.J., Waltman, A.W., Day, M.W. and Grubbs, R.H. (2003) Angew. Chem., 115, 3403–3407; Angew. Chem. Int. Ed., 42, 3281–3285; (b) Guidry, E.N., Cantrill, S. J., Stoddart, J.F. and Grubbs, R.H. (2005) Org. Lett., 7, 2129–2132. 51 Iwamoto, H., Itoh, K., Nagamiya, H. and Fukazawa, Y. (2003) Tetrahedron Lett., 44, 5773–5776. 52 Kidd, T.J., Leigh, D.A. and Wilson, A.J. (1999) J. Am. Chem. Soc., 121, 1599– 1600. 53 (a) Vysotsky, M.O., Bolte, M., Thondorf, I. and Böhmer, V. (2003) Chem. Eur. J., 9, 3375–3382; (b) Wang, L., Vysotsky, M.O., Bogdan, A., Bolte, M. and Böhmer, V. (2004) Science, 304, 1312–1314; (c) Bogdan, A., Vysotsky, M.O., Ikai, T., Okamoto, Y. and Böhmer, V. (2004) Chem. Eur. J., 10, 3324–3330. 54 Molokanova, O., Vysotsky, M.O., Cao, Y., Thondorf, I. and Böhmer, V. (2006)

References

55

56

57

58

59

60

61

62

63

64

65

Angew. Chem., 118, 8220–8224; Angew. Chem. Int. Ed., 45, 8051–8055. (a) Wisner, J.A., Beer, P.D., Drew, M.G. B. and Sambrook, M.R. (2002) J. Am. Chem. Soc., 124, 12469–12476; (b) Sambrook, M.R., Beer, P.D., Lankshear, M.D., Ludlow, R.F. and Wisner, J.A. (2006) Org. Biomol. Chem., 4, 1529–1538. (a) Sambrook, M.R., Beer, P.D., Wisner, J.A., Paul, R.L. and Cowley, A.R. (2004) J. Am. Chem. Soc., 126, 15364–15365; (b) Ng, K.-Y., Cowley, A.R. and Beer, P.D. (2006) Chem. Commun., 3676–3678. Borisova, N.E., Reshetova, M.D. and Ustynyuk, Y.A. (2007) Chem. Rev., 107, 46–79. Hutin, M., Schalley, C.A., Bernardinelli, G. and Nitschke, J.R. (2006) Chem. Eur. J., 12, 4069–4076. Ziessel, R., Harriman, A., Suffert, J., Youinou, M.-T., De Cian, A. and Fischer, J. (1997) Angew. Chem., 109, 2621–2624; Angew. Chem. Int. Ed., 36, 2509–2511. (a) Glink, P.T., Oliva, A.I., Stoddart, J.F., White, A.J.P. and Williams, D.J. (2001) Angew. Chem., 113, 1922–1927; Angew. Chem. Int. Ed., 40, 1870–1875; (b) Horn, M., Ihringer, J., Glink, P.T. and Stoddart, J.F. (2003) Chem. Eur. J., 9, 4046–4054. (a) Chichak, K.S., Cantrill, S.J., Pease, A. R., Chiu, S.-H., Cave, G.W.V., Atwood, J. L. and Stoddart, J.F. (2004) Science, 304, 1308–1312; (b) Chichak, K.S., Cantrill, S. J. and Stoddart, J.F. (2005) Chem. Commun., 3391–3393. Peters, A.J., Chichak, K.S., Cantrill, S.J. and Stoddart, J.F. (2005) Chem. Commun., 3394–3396. Pentecost, C.D., Chichak, K.S., Peters, A.J., Cave, G.W.V., Cantrill, S.J. and Stoddart, J.F. (2007) Angew. Chem., 119, 222–226; Angew. Chem. Int. Ed., 46, 218– 222. Fujita, M., Ibukuro, F., Hagihara, H. and Ogura, K. (1994) Nature, 367, 720–723. Fujita, M., Ibukuro, F., Yamaguchi, K. and Ogura, K. (1995) J. Am. Chem. Soc., 117, 4175–4176.

66 Fujita, M., Fujita, N., Ogura, K. and Yamaguchi, K. (1999) Nature, 400, 52–55. 67 Yamashita, K.-i., Kawano, M. and Fujita, M. (2007) J. Am. Chem. Soc., 129, 1850–1851. 68 Ibukuro, F., Fujita, M., Yamaguchi, K. and Sauvage, J.-P. (1999) J. Am. Chem. Soc., 121, 11014–11015. 69 (a) Dietrich-Buchecker, C.O., Geum, N., Hori, A., Fujita, M., Sakamoto, S., Yamaguchi, K. and Sauvage, J.-P. (2001) Chem. Commun., 1182–1183; (b) Dietrich-Buchecker C.O., Colasson, B., Fujita, M., Hori, A., Geum, N., Sakamoto, S., Yamaguchi, K. and Sauvage, J.-P. (2003) J. Am. Chem. Soc., 125, 5717–5725. 70 (a) Chang, S.-Y., Jang, H.-Y. and Jeong, K.-S. (2003) Chem. Eur. J., 9, 1535–1541; (b) Chang, S.-Y. and Jeong, K.-S. (2003) J. Org. Chem., 68, 4014–4019. 71 (a) Heim, C., Affeld, A., Nieger, M. and Vögtle, F. (1999) Helv. Chim. Acta, 82, 746–759; (b) Parham, A.H., Windisch, B. and Vögtle, F. (1999) Eur. J. Org. Chem., 1233–1238; (c) Seel, C., Parham, A.H., Safarowsky, O., Hübner, G.M. and Vögtle, F. (1999) J. Org. Chem., 64, 7236–7242. 72 Chas, M., Pia, E., Toba, R., Peinador, C. and Quintela, J.M. (2006) Inorg. Chem., 45, 6117–6119. 73 Blight, B.A., Wisner, J.A. and Jennings, M.C. (2007) Angew. Chem., 119, 2893–2896; Angew. Chem. Int. Ed., 46, 2835–2838. 74 (a) Hunter, C., (1991) J. Chem. Soc., Chem Commun., 749–751; (b) Hunter, C. and Purvis, D.H. (1992) Angew. Chem., 104, 779–782; Angew. Chem., Int. Ed., 31, 792–795. 75 Blight, B.A., Wisner, J.A. and Jennings, M.C. (2006) Chem. Commun., 4583– 4585 76 Blight, B.A., Van Noortwyk, K.A., Wisner, J.A. and Jennings, M.C. (2005) Angew. Chem., 117, 1523–1528; Angew. Chem. Int. Ed., 44, 1499– 1504.

j391

j393

Index a 16-membered ring 44, 47 6-aminopicolinic acid 334 [12]annulenes 211 [14]annulenes 193, 210 [15]annulenes 193, 210 [18]annulenes 186, 21 activated urethane 148 adamantane-2,6-dione 101 D-Ala-D-Ala peptide motif 5 alkane oxidation 306 alkene metathesis 351, 369, 372, 386 alkyne metathesis 186, 213 allosteric group 123 allosteric interactions 336 amino acid 297 amphidinolide A 37 analytical chemistry 315 anion template 373, 375 anion-templated assembly 316 anion transport 318 anisotropic LC 216 anti-cancer agents 123 apical functionalization approach 236 Arg-Gly-Asp (RGD) tripeptide 17 aromatic spacers 279 aromatic stacking 82 aryl–perﬂuoroaryl interactions 209 aryltriazene 191 Aspicillin 40 association constants 350 atropdiastereomers 46

b

4,40 -bipyridine 191, 195, 331 barbiturate 78 Behrend's polymer 113 binding epitopes 122 biology 315

bio-supramolecular chemistry 22 biphenyl spacer 332 bis[2]catenane(s) 144, 155, 163, 164, 165, 168 – regioisomeric 164 bis[2]rotaxanes 170 bis[3]catenane(s) 163, 168, 169 bis-crown-3 151 bisloop calix[4]arena(s) 156, 163 bisloop compound(s) 158, 162, 163, 170, 174 – ESI mass spectra 158 – 1H NMR spectra 158 bisloop tetra-urea(s) 157, 170, 172, 174 bis(paraquat-p-phenylene) 147 bis-p-phenylene-[34]crown-10 176 bis-rosette motifs 76 Boc-groups 149 Borromean rings 143, 144, 386, 378, 379 Breslow coupling 186 206, 225 bridging units 248 – picolyl groups 248 bromination 204 tert-butylcalix[4]arene 148

c [2]catenane(s) 144, 199 [5]catenane (olympiadane) 145 C2v-symmetric calixarene 152 C4-symmetry, calixarene 152 Cadiot–Chodkiewicz coupling 190, 210, 224 – Pd-catalyzed 224 – homocoupling 193 cage self-assembly (CSA) 239 – factors controlling 239 calix[4]arene(s) 97, 155, 373 – multimacrocyclic 155 – wide rim 155 calixarenes 90, 223, 317 calix-bisporphyrin 237 calix-tetraporphyrin 237

Modern Supramolecular Chemistry: Strategies for Macrocycle Synthesis. Edited by François Diederich, Peter J. Stang, and Rik R. Tykwinski Copyright Ó 2008 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim ISBN: 978-3-527-31826-1

j Index

394

capsules 90 carbonate link 144 carbon dioxide 336 Castro–Stephens coupling 188, 206 catalysis 96 catenane(s) 59, 60, 143, 144, 145, 146, 163, 168, 176, 339 –p-conjugated 199 cation–p interactions 150, 151 cavitand 90, 223, 245 – octadentate 245 cavitand complex 265 – cis-ditopic 265 – trans-ditopic 266 – cis-ditopic neutral 267 – tetratopic 266 cavitand ligand 264 – monodentate 264 cavities 90 CB[6] 122 CB[7] 124 CB[8] 123 CB[10]CB[5] 117 CB[n] homologs 115 charge transfer interaction 120 chelate effect 76 chelates, 5-membered 35 chelates, 6-membered 35 Chiralcel OD 165 chiral memory 80 Chiralpak AD-H 165 chiral stationary phases 165 cholapods 317 chromatographic resolution 166 chromatographic separation 327 cleft-type receptors 223 clipping 145, 170, 171, 350, 354, 383, 384, 376 – methodology 363 – reaction 171 cobaltocenium 150 coleophomones 48, 58 combinatorial synthetic methods 341 commercial product 278 compartmentalization 234 complexes 278, 280 – hollow 278 – M6L4 octahedral 280 condensation 354 cone conformation 148, 156, 160 conformational interconversion 160 a-connection(s) 154, 155, 164 b-connection(s) 154, 155, 164, 165 b0 -connection 164, 165, 167 cooperative binding 320, 336

cooperativity effects 321 coordination cages 234, 235, 252 – cavitand-based dimeric 235 – dimeric calixarene-based 235 – hexameric 252 – self-assembled M6L4 277 – tetrameric 252, 254 – trimeric 252, 254, 256 copper acetylide 364 covalent prearrangement 144 crown ether(s) 145, 147, 317 crown ether system 319 – 19-benzocrown-6 320 – dibenzo-24-crown-8 macrocycle 319 cryptands 332, 324 – N7-azacalix[3]cryptand 324 crystal structure 134 CsA's biological activity 12 C-shaped forms 126 CtA 15 – total synthesis 15 cucurbit[n]uril 113 – analogs 126 – derivatives 129 – mechanism 113 cyanuric acid 76 cyclene 319, 320 cyclene-dibenzacrown hybrid macrocycle 319 cyclic [8]catenane(s) 163, 168, 169 cyclic anhydride structures 169 cyclic peptides with SH2-domains 2 cyclobis(paraquat-p-phenylene) 176 cyclodextrin 116 – homodimers 61 cyclopeptide 44, 334, 335 – bis(cyclopeptides) 335 cyclophane 333 cycloproparadicicol 57 cyclosporin 1 cyclosporin A 2, 10 – immunosuppressive activity 10 cyclotheonamide 2, 13 – total synthesis 14, 15 cyclotriveratrylene 90 cylindrocyclophanes 47 cytotrienins 37

d D2-symmetry 152 1,4-diazabicyclo[2.2.2]octane (DABCO) 331 2,6-diiminopyridine 144 3,4-diiodothiophene 189 3,5-diethynylpyridine 195 dehydrobromination 204

Index dendrimer(s) 83, 173, 119 – self-assembly 173 – structurally uniform 175 dendritic assembly 173 dialkylammonium 372, 377 dibenzo crown 145 Diels–Alder cycloaddition 170 Diels–Alder reaction 303 diffusion-ordered 1H NMR spectroscopy (DOSY) 85, 173 dimeric capsule(s) 106, 149 – expansion 106 dinuclear Pt precursors 243 diol protecting groups 39 dipolar cycloaddition 114 direct functionalization 131 discotic liquid crystal 209 discrete siloxane synthesis 308 disruption of H-bonding 94 dissociation 94 dissymmetric 98 dithiocarbamate (dtc) motif 331 donor-acceptor interactions 176 double-helical template 351 double rosettes 78 double template synthesis 160 doubly bridged [2]catenane 155 drug-delivery devices 91 dumbbell shaped molecule 145, 147 dynamic combinatorial chemistry 316

e echistatin 19 effective molarity 70 Eglinton coupling 190, 193, 205, 210, 225, 364 encapsulation 150 – selectivity 247 ene-yne metathesis 45 enthalpic gain 75 enthalpy–entropy compensation 69 environmental remediation 315 epothilone 34 epothilone A 44 eremomycin 6 ESI mass spectra 167, 169, 170 exchange rate, guest 103 – process 89 experimental 22, 106, 177, 224, 271, 386 – [2]catenane 14 386 – [2]catenane 25 387 – [2]catenane 118 387 – [2]rotaxane 81 387 – [8]catenane 14 179

– AC-dibridged tolylpyridyl cavitand 35 272 – alkene-bridged mimic 22 – vancomycin 22 – alkyne-bridged mimic 22 – nisin 22 – bis[2]catenane 12 177 – bisloop tetra-urea 8 177 – crystals of the host–guest arrangement of 52@(50b)6 107 – cucurbit[n]urils 138 – fac-Br(CO)3Re AC ditopic cavitand complex 36 272 – glycolurils 137 – hexameric capsule derived from pyrogallol [4]arene (50c) 106 – large-scale synthesis of SPM 124 226 – macrocycle 55 342 – macrocycle H2SO453 342 – pre-catenane 56 225 – SPM 13 224 – SPM 19 224 – SPM 20 225 – SPM 33 225 – SPM 91 226 – synthesis of M6L4 cage 2 309 – tetracyano cavitand 40 272 – tetraloop tetra-urea 9 178 – tetrapicolyl-bridged cavitand 31a 271 – tetratopic cavitand complex 48 273 – tetra-urea 5 177 – tetra-urea 6 178 extraction 338 E-Z photoisomerization 82 E/Z selectivity 39, 44, 46, 51

f ferrocene receptor 320 – amidoferrocene 320 ﬂuvirucin 37 fourfold [2]rotaxanes 171 free-radical polymerization 199 fullerene dimers 83

g gel permeation chromatography 80, 83 generation of 3D architectures 234 giant macrocycles 160 Glaser coupling 364 gloesporone 36 b-D-glucoside 326 glycoluril(s) 99, 115 glycouril spacer 106 glyoxal solution 137 gold 258

j395

j Index

396

G-quartet 86 guanidiniums 316 guest exchange 165, 171 guest inclusion 153

h halide template 373 Hay coupling 195, 196, 205, 364 head-to-head dimer 33 head-to-tail manner 99 helicate 351 herbarumin II 41 heterodimeric capsule 171 heterodimerization 129, 158, 172 heterodimers 151, 155, 156, 170, 172, 173 – 1H NMR spectra 162 heteroleptic complex 199 hexamer, resorcinarene 103 hexameric capsule 101 hierarchical self-assembly 83 high-dilution conditions 350 1 H NMR spectra 149, 167, 168 homodimer(s) 151, 156, 170, 172 – 1H NMR spectra 162 homoleptic complex 199 homotropic allostery 135 huge macrocycles 161 hydrogenation 369, 373 hydrogen bonded belt 149, 150, 162, 168 hydrogen-bonded dimers 169 hydrogen-bonded template 372 hydrogen bonding 145 – template 363 hydrogen bonds 145, 153 – directionality 153 hydrophobic binding 118 hydroquinone 147

i imidazoliums 316 imine metathesis 214 iminium ion 126 immunosuppressants 12 inclusion 288 – adamantane 288 – carborane 288 – of two guest molecules 171 inclusion complex 293 – heterotopic 293 inclusion geometry 289 integrins 17 –aVb3 17 – structural basis for binding 17 p-p interactions 120, 147

interlocked molecules 349 interlocked rings 143 interlocked structures 144 internal volume 172 intramolecular connection 154 inverted CB[n] 130 ion channels 132 ion-dipole interactions 114 ion pairing 82, 373, 205 ion-selective electrodes 327 ion sensor 205 irradiation 381 isocyanate 148 isonicotinate 331, 332 isophthalic acid 145 isothermal titration calorimetry (ITC) 319

j Janus type molecules 80 josamycin 47

k a-ketoarginine 13 knots 143, 168

l lactam, 14-membered, ﬂuvirucin 31, 32 lantibiotic(s) 2, 6, 9 larger assemblies, calixarenes 172 large-scale production, coordination cage 280 lectins 132 light scattering 173 limiting reagent 363 lipid II 5, 8, 10 – pore formation 9 – pyrophosphate moiety 9 – speciﬁc nisin activity 9 longithorone A 45 luminophor 206

m macrocycle aggregation 208 macrocycle ring-closing metathesis (M-RCM) 29 macrocycle(s) 29, 50, 154, 342 macrocyclic alkene 42 macrocyclic peptide structures 22 macrolactams 29 macrolactones 29 MALDI-TOF-MS 170 McMurry reaction 204 MD simulation 171 mechanical bond 349

Index medicine 315 melamine 76 memory effect 98 Menschutkin reaction 351 metal-based templation 59 metal-carbene precatlysts 30 metal ion coupling 205 metal-templated guanine 87 metal to ligand charge transfer (MLCT) 341 metathesis 29, 156, 161, 170 – reaction 144 methylene-bridged glycoluril dimers 115 microcarpalide 41 Möbius strips 143 molecular container(s) 91, 114 molecular gyroscopes 60 molecular ice 296 molecular knot(s) 59, 60 molecular lock 378 molecular loop lock 120 molecular machines 119 molecular mechanisms 22 – interaction 22 molecular recognition 134 molecular shuttle 115 molydenum catalyst 33 molybdenum precatalyst 45 monocillin I 56 monoloop compounds 172 monoloop tetra-urea compounds 167 b-monosaccharides 336 multicatenanes 173, 176 multicomponent reactions (MCR) 50 multimacrocycles 155 multiple catenanes 154, 162, 164 multiple rotaxanes 170 multi-rotaxanes 176 multitopic 69 multitopic macrocyclic complexes 263 multivalent 132 muscone 33, 50

n nanobasket 199 nanoﬁbrils 213 nisin 6, 8 nisin–lipid II complex 9 ipso-nitration 148 NMR timescale 89 novel topologies 148, 166 nuclear waste remediation 338 nucleotides 328 – analogs 328 – diphosphates 328

o O alkylation 148 octadentate bipyridine cavitand 246 octahedral coordination 144 oleﬁn metathesis 145, 154, 169 oligomer(s) 69 olympiadane 354 one-pot approach 188 onion-type complexes 205 optical resolution 165 orthogonal alkyne protecting groups 192 oxidative acetylenic homocoupling 186, 219, 365, 369 – Eglinton coupling 186 – Glaser coupling 186 – Hay coupling 186 – in situ deprotection 219 oxidative coupling 364 oximidines 51, 55

p 2,6-pyridine diacid 145 2,6-pyridine dialdehyde 144 pair-wise dimers 96 paracyclophane 47 paraphenylacetylene macrocycles 204 paraquat 351 partial cone conformation 156, 160 peptide sequence 297 peripheral crowding 78 peripheral stabilization 82 phenanthroline ligand 199 a-phenylglycine anions 322 phenyl spacer 250 phospholipid bilayer membrane 318 photoaddition 303 photodimerization 302 photoswitchable [2]catenane 381 phthalhydrazides 129 pincer–pyridine complexes 60 platinum diacetylide 203, 364 polar complexing groups 35 polyammoniums 316 polymers 83 pore formation 9 – model 9 porous solid 196 porphyrin 191, 195, 330 – tetraphenylporphyrin scaffold 330 porphyrin oligomers 76 portals 94 prearrangement 143 preorganization 74, 75, 145, 147, 148, 154, 155, 176

j397

j Index

398

– functional groups 170 L-proline 334 protecting-group effect 53 protection–deprotection 148 protein 297 pseudo-high dilution 186, 209 pseudorotaxane(s) 146, 170, 176, 339 pyridinophanes 210 pyrogallol[4]arenas 103 pyrophosphate binders 10 pyrroles 316

ring-opening metathesis polymerization (ROMP) 199 ristocetin A 6 robust self-assembled macrocycles 89 rosette system 74 rotaxane 341 rotaxane(s) 59, 144, 145, 146, 168, 173, 176, 339 – axle 176 – constitutionally isomeric 176 – wheel 176 ruthenium catalysts 30, 31

q quadruplexes 86 quadrupole–quadrupole interactions 210 quaternization 351, 358

s

salen-based anion 332 salicylihalamides 51 sapphyrin 328 r – oligosapphyrins 328 Schiff base 144, 373 racemization 98 Schiff base condensation 207 radicicol 56 Schrock's molybdenum catalysts 31, 47 RCM/hydrogenation 49 secondary dialkylammonium 372 RCM/ROM/RCM cascade 48 secondary interaction 70 recifeiolide 42 second-sphere coordination 386 recognition 291, 293 self-anchored rotaxane 176 – azobenzene 295 self-assembled capsules 153 – bimolecular 291 self-assembled dimers 176 – bulky guests 293 self-assembly 175 – peptide 297 self-assembly, quadrupole enhanced 209 – stilbene 295 self-organization 173 reductive elimination 364 self-recognition 384 regioisomeric bis[2]catenanes 163 self-sorting 121, 172, 174 relay ring-closing metathesis 55 – process 173 resorcinarene(s) 90, 103, 104 separation factors 165 reversible [2þ2]-cycloaddition 32 SH2 domain-binding analogs 20 RGD 18, 19 SH2 domain-binding peptides 19 RGD-containing cyclic peptides 2 SH2 domains 19, 21 RGD peptides 16 shape persistent macrocycle 44 rigidifed tetra-ureas 173 – adsorption on surfaces 218 ring-closing alkene metathesis 10 – aggregation 208 – alkane mimics 10 – arylene–ethynylene 205 – alkene mimics 10 – Au(1 1 1) surface 211 – alkyne mimics 10 ring-closing alkyne metathesis (RCAM) 10, 45 – binaphthyl-derived 203 – C60 inclusion 204 – alkane mimics 10 – C70 inclusion 204 – alkene mimics 10 – alkyne mimics 10 – carbazole-based 213 ring-closure 144 – carbohydrate recognition 204 – reactions 170 – catenanes 199 ring closing metathesis (RCM) 29, 341, 369, – columnar structure 216 370, 371, 372, 373 – complex with C60 220 ring closing reaction 169 – discotic LC 217 ring contraction 47 – face-to-face aggregation 210 ring-expansion metathesis 48 – ﬂuorescence 196, 206, 208

Index – helicene-based 211 – honeycomb structure 220 – host–guest systems 203 – hydrogen-bonded 203 – inverted LC 217 – Kagomé lattice 220 – liquid crystals (LC) 215 – metal iron coupling 205 – oligothiophene-based 203, 222 – on an HOPG surface 220 – optimized synthesis 199 – ortho-phenylene ethynylene 216 – polymerization 214 – Pt-acetylide 203 – pyridinium 212 – scanning tunneling microscopy (STM) analysis 220, 218, 222 – schiff base 208 – self aggregation 211 – self-assembly 195, 199 – solid support 193 – surface chemistry 208 – synthesis 186 – templated synthesis 190, 207 – terpyridine based 22 – thin ﬁlms 214 – tubular structures 212 – vapor pressure osmometry (VPO) 212 – X-ray crystallographic analysis 195, 197, 206, 218, 224 – X-ray diffraction analysis 215 single-crystal X-ray analysis 168 soft ball, molecular 91 sol–gel method 213 solid-liquid transport scenario 318 solid support 193 Solomon knot 351, 378, 379, 381, 382 Sonogashira coupling 186, 188, 190, 191, 192, 196, 199, 203, 206, 209, 215 spacers, resorcinarene 241 spectroﬂuorometric 104 spin–spin interaction 299 – intermolecular 299 S-shaped diastereomers 126 p–p stacking 145 steric hindrance 75 steroid(s) 316, 325, 326 Stille coupling 199 stoppering 170, 171 stoppers 145, 176 strapped calixarenes 322 strapped diazacrown complex 318 strap(s) 317, 322, 328

– cis-strapped 328 substrate–receptor compatibility 315 sulfonamides 99 supramolecular polymer 118 supramolecular structures 176 surfaces 255 – self-assembly of coordination cages 255 switchable polymers 172 symmetry classes 153 symmetry elements 153 symmetry properties 152, 153, 165 synthesis of bis[2]catenanes 163, 164

t tag, peptide 297 tandem CM/RCM 47 tandem RCM/ROM/RCM 48 tandem ROM/RCM 47 teicoplanin 6 template-based assembly 317 template-directed RCM 37 templating cation 144 templating group 116 templating metal 89 template(s) 59, 145, 156, 161, 171, 350 – effect 145 – synthesis 156 tennis ball 91 terephthalate 331 ternary complex 119 terpyridyl 144 tetraamino calix[4]arene 161 tetra-aryl-tosylureas 173 tetra-bridged capsule 155 tetradentate cavitands 243 tetraethylsilane ferrocene 151 tetraloop calix[4]arena(s) 156, 171 tetraloop compound(s) 156, 158, 160, 162, 163, 170 – ESI mass spectra 158 – single-crystal X-ray structure 160 tetraloop compounds 9 160 tetraloop tetra-urea(s) 157, 170, 172 tetra[2]rotaxane 170 tetrathiafulvalene (TTF) 211 tetra-tosylurea 4 156 tetra-tosylurea(s) 161, 171, 173 tetra-urea calyx[4]arenes 148, 149, 151, 153, 155, 176 – dimers 153 – preorganization 153 tetra-urea(s) 148, 151, 173 thermodynamic control 40, 186

j399

j Index

400

thermodynamic versus kinetic issue 32 thermotropic mesophases 218 thioether bridge(s) 10 thiourea NH proton signals 324 threading 176, 358 thrombin 13 time-averaged conformation 160 titanium isopropoxide 37, 48, 50 p-toluenesulfonic acid 124 tonantzitlolone 43 topologically chiral 168 topology 168 trefoil knot(s) 144, 145, 351, 355 triphenylmethanes 173 triptycene core 145 trisloop compound(s) 156, 162, 163 trisloop derivatives 164 trisloop tetra-urea(s) 157, 170 tritylchloride 149 trityl-groups 149 tri-ureas 26 173

v vancomycin 1, 4, 6 – mimics 6 – resistance 6 – total syntheses 6 vancomycin aglycon 5 vapor pressure osmometry 80, 85, 212 vinylogous L-tyrosine 13 viologen 119 viscosity 83

w Wacker oxidation 306 wide rim 176 Williamson ether synthesis 351

x 129

Xe NMR spectroscopy 196 X-ray analysis 165, 166, 195, 197, 206, 218, 224 X-ray diffraction 150, 215 X-ray structure 149

u Ugi 4CC-IMDA sequence 39 ultrahigh pressure 354 urea 137

y ynolide 57